alyferryhalo/cifar10_cnn_homework_ver1-0.ipynb

## cifar10_cnn_homework_ver1-0.ipynb
{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "cifar10_CNN_homework_ver1-0.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "ToqeRcFqnzR9"
      ],
      "authorship_tag": "ABX9TyOwoPociaqjzfOGAkWjaku3",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/gist/alyferryhalo/c327625e63f05dedfeb9b8918c391483/cifar10_cnn_homework_ver1-0.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IutcvtVEmkPz"
      },
      "source": [
        "# Классификация изображений cifar10\n",
        "\n",
        "В этом ноутбуке попробуем разные модели и гиперпараметры для получения наилучшего результата на датасете cifar10.\n",
        "\n",
        "Данный ноутбук — часть домашнего задания по курсу \"Нейросетевые технологии в компьютерном зрении\" от пермского сетевого IT-университета."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fEtyW2FemugJ"
      },
      "source": [
        "## Отображатель графиков обучения"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mbOlwMejmhUI"
      },
      "source": [
        "from matplotlib import pyplot as plt\n",
        "from tensorflow.keras.callbacks import Callback\n",
        "from IPython.display import clear_output\n",
        "from tensorflow.keras import backend as K\n",
        "\n",
        "\n",
        "class PlotLearning(Callback):\n",
        "    def on_train_begin(self, logs={}):\n",
        "        self.i = 0\n",
        "        self.x = []\n",
        "        self.losses = []\n",
        "        self.val_losses = []\n",
        "        self.inter_dim = []\n",
        "        self.val_inter_dim = []\n",
        "        self.logits = []\n",
        "        self.val_logits = []\n",
        "        self.acc = []\n",
        "        self.val_acc = []\n",
        "        self.fig = plt.figure()\n",
        "\n",
        "        self.logs = []\n",
        "\n",
        "    def on_epoch_end(self, epoch, logs={}):\n",
        "\n",
        "        self.logs.append(logs)\n",
        "        self.x.append(self.i)\n",
        "        self.losses.append(logs.get('loss'))\n",
        "        self.val_losses.append(logs.get('val_loss'))        \n",
        "        self.acc.append(logs.get('accuracy'))\n",
        "        self.val_acc.append(logs.get('val_accuracy'))\n",
        "\n",
        "        self.i += 1\n",
        "        f, (ax1, ax2) = plt.subplots(2, 1, sharex=True)\n",
        "\n",
        "        clear_output(wait=True)\n",
        "\n",
        "        ax1.set_yscale('log')\n",
        "        ax1.plot(self.x, self.losses, label=\"loss:\" +\n",
        "                 str(round(logs.get('loss'), 2)))\n",
        "        ax1.plot(self.x, self.val_losses, label=\"v_loss:\" +\n",
        "                 str(round(logs.get('val_loss'), 2)))\n",
        "        ax1.legend()\n",
        "        \n",
        "        ax2.plot(self.x, self.acc, label=\"acc:\" +\n",
        "                 str(round(logs.get('accuracy'), 2)))\n",
        "        ax2.plot(self.x, self.val_acc, label=\"v_acc:\" +\n",
        "                 str(round(logs.get('val_accuracy'), 2)))\n",
        "        ax2.legend()\n",
        "\n",
        "        plt.savefig(\"fig\")\n",
        "        plt.show()\n",
        "\n",
        "\n",
        "plot = PlotLearning()"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "K2B_-a2Lm14L"
      },
      "source": [
        "## Загрузка и исследование датасета"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 352
        },
        "id": "I-bV7Oibm7Sl",
        "outputId": "6dd2aedf-1ca7-4457-e165-cce0874a8462"
      },
      "source": [
        "from __future__ import print_function\n",
        "from tensorflow.keras.datasets import cifar10\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Dense, Dropout, Activation, Flatten, BatchNormalization\n",
        "from tensorflow.keras.layers import Conv2D, MaxPooling2D, AveragePooling2D\n",
        "import os\n",
        "import tensorflow.keras as keras\n",
        "import numpy as np\n",
        "\n",
        "num_classes = 10\n",
        "\n",
        "save_dir = os.path.join(os.getcwd(), 'saved_models')\n",
        "model_name = 'keras_cifar10_trained_model.h5'\n",
        "\n",
        "(x_train, y_train), (x_test, y_test) = cifar10.load_data()\n",
        "\n",
        "plt.hist(y_train, num_classes)\n",
        "plt.show()\n",
        "\n",
        "y_train = keras.utils.to_categorical(y_train, num_classes)\n",
        "y_test = keras.utils.to_categorical(y_test, num_classes)\n",
        "x_train = x_train.astype('float32')\n",
        "x_test = x_test.astype('float32')\n",
        "print('x_train shape:', x_train.shape)\n",
        "print(x_train.shape[0], 'train samples')\n",
        "print(x_test.shape[0], 'test samples')\n",
        "x_train /= 255\n",
        "x_test /= 255"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading data from https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz\n",
            "170500096/170498071 [==============================] - 4s 0us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "x_train shape: (50000, 32, 32, 3)\n",
            "50000 train samples\n",
            "10000 test samples\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3V2Qev7Tm_-L"
      },
      "source": [
        "## Создание и обучение тестовой модели №1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_M6OQPX0nBa5",
        "outputId": "330fecb2-6472-4188-99ef-dd3d0d61ea98"
      },
      "source": [
        "from tensorflow.keras.layers import Dropout, BatchNormalization, LeakyReLU, PReLU\n",
        "\n",
        "batch_size = 128\n",
        "epochs = 300\n",
        "\n",
        "model = Sequential()\n",
        "model.add(Conv2D(16, (3,3), input_shape=x_train.shape[1:], kernel_initializer='he_normal'))\n",
        "model.add(LeakyReLU(0.1))\n",
        "model.add(BatchNormalization())\n",
        "model.add(MaxPooling2D(2))\n",
        "\n",
        "model.add(Conv2D(32, (3,3)))\n",
        "model.add(LeakyReLU(0.1))\n",
        "model.add(BatchNormalization())\n",
        "model.add(MaxPooling2D(2))\n",
        "\n",
        "model.add(Flatten())\n",
        "model.add(Dropout(0.3))\n",
        "\n",
        "model.add(Dense(32, kernel_initializer='he_normal'))\n",
        "model.add(LeakyReLU(0.1))\n",
        "\n",
        "model.add(Dense(num_classes, kernel_initializer='he_normal'))\n",
        "model.add(Activation('softmax'))\n",
        "\n",
        "# initiate optimizer\n",
        "opt = keras.optimizers.Adam()\n",
        "\n",
        "# Let's train the model using RMSprop\n",
        "model.compile(loss='categorical_crossentropy',\n",
        "              optimizer=opt,\n",
        "              metrics=['accuracy'])\n",
        "model.summary()"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d (Conv2D)              (None, 30, 30, 16)        448       \n",
            "_________________________________________________________________\n",
            "leaky_re_lu (LeakyReLU)      (None, 30, 30, 16)        0         \n",
            "_________________________________________________________________\n",
            "batch_normalization (BatchNo (None, 30, 30, 16)        64        \n",
            "_________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D) (None, 15, 15, 16)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_1 (Conv2D)            (None, 13, 13, 32)        4640      \n",
            "_________________________________________________________________\n",
            "leaky_re_lu_1 (LeakyReLU)    (None, 13, 13, 32)        0         \n",
            "_________________________________________________________________\n",
            "batch_normalization_1 (Batch (None, 13, 13, 32)        128       \n",
            "_________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2 (None, 6, 6, 32)          0         \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 1152)              0         \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 1152)              0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 32)                36896     \n",
            "_________________________________________________________________\n",
            "leaky_re_lu_2 (LeakyReLU)    (None, 32)                0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 10)                330       \n",
            "_________________________________________________________________\n",
            "activation (Activation)      (None, 10)                0         \n",
            "=================================================================\n",
            "Total params: 42,506\n",
            "Trainable params: 42,410\n",
            "Non-trainable params: 96\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "qFDorCFEni5f",
        "outputId": "a601e951-a94a-49e7-c611-a1f36593dddc"
      },
      "source": [
        "print('Not using data augmentation.')\n",
        "\n",
        "model.fit(x_train, y_train,\n",
        "          batch_size=batch_size,\n",
        "          epochs=epochs,\n",
        "          validation_data=(x_test, y_test),\n",
        "          shuffle=True,\n",
        "          callbacks=[plot]\n",
        "          )"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7ff93fc764d0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ToqeRcFqnzR9"
      },
      "source": [
        "### Отображение результатов"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "TLY6sQfkn30U",
        "outputId": "6f59c63b-d89d-469c-d7eb-656a3a140b75"
      },
      "source": [
        "from matplotlib.colors import Colormap\n",
        "from mpl_toolkits.axes_grid1 import ImageGrid\n",
        "import numpy as np\n",
        "import random\n",
        "\n",
        "def get_class_sample(cls_index):\n",
        "    for i in range(len(y_train)):        \n",
        "        if np.argmax(y_train[i]) == cls_index:\n",
        "            return x_train[i]\n",
        "    return None\n",
        "\n",
        "print(\"          GT class           Predicted class\")\n",
        "\n",
        "for i in range(10):    \n",
        "    plt.set_cmap(Colormap(\"Greys\"))\n",
        "    fig = plt.figure(1, (4., 4.))\n",
        "    grid = ImageGrid(fig, 111, \n",
        "                 nrows_ncols=(1, 2), \n",
        "                 axes_pad=0.1,\n",
        "                 )\n",
        "\n",
        "    images = []\n",
        "    np_images = []\n",
        "    k = random.randint(0,len(x_test)-1)\n",
        "    images.append(np.expand_dims(x_test[k], axis=0))\n",
        "    np_images = np.vstack(images)\n",
        "    y = model.predict(np_images)\n",
        "    \n",
        "    c = np.argmax(y[0])\n",
        "    \n",
        "    cls_name_pred = c\n",
        "    cls_name_gt = np.argmax(y_test[k])\n",
        "    \n",
        "    img = x_test[k]\n",
        "    img_pred = get_class_sample(cls_name_pred)\n",
        "    grid[0].imshow(np.reshape(img,(32,32,3)))\n",
        "    grid[0].set_title(\"{}\".format(cls_name_gt))\n",
        "    grid[1].imshow(np.reshape(img_pred,(32,32,3)))\n",
        "    grid[1].set_title(\"{}\".format(cls_name_pred))\n",
        "\n",
        "    plt.show()\n",
        "    "
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "          GT class           Predicted class\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAADTCAYAAACY5kjLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO2de3Bd13nd13cfuHiRBAHwTYoPUZFCSdYjHL8kOZIfreM2keRxOkkmHjfxVJnOeCYZ548ojzYP5w8nTWy3du1UqZ0orRrHieWx4iiNZUet61hWTEkWJZGSSFGiJIpPECDewH3s/oGrFsBaR7wgSADHWb8ZDMEP556zzz77bByctdf3RUoJxhhj8kdhuRtgjDHmwvAEbowxOcUTuDHG5BRP4MYYk1M8gRtjTE7xBG6MMTnFE7gxxuQUT+A5ISI+EhH7ImIqIv50udtjTBYeq0tHabkbYFrmNQC/C+CfA+hY5rYY80Z4rC4RnsBzQkrpfgCIiL0Ati5zc4zJxGN16fArFGOMySmewI0xJqd4AjfGmJziCdwYY3KKRcycEBElzFyvIoBiRLQDqKWUasvbMmPm4rG6dPgJPD/8BoAJAHcD+Nnm97+xrC0yRuOxukSECzoYY0w+8RO4McbkFE/gxhiTUzyBG2NMTvEEbowxOWVRE3hEvDcinouIwxFx98VqlDHGmPNzwatQIqII4HkA7wHwKoDvAfjplNKBrM/09vWnrdu2X9DxMlqREW/tnOSns3Z5odtlNCXEDuSm4vrI7VR7xIatNnsh5HUd01NPPn4mpbRO/Wz1mp60bv2meVE+0wh+BioU1LXVz0qq79TYiJavpdqwtbGWdR1DzhFin2Kzlseq4hKskFvMHnVzWt9jy30uDnT0hQNyrC7GyPNmAIdTSkcAICK+COB2AJkT+NZt2/Hg3z9ygYfjq95AMWPLBscSxwpiu0JR3GjihlSDUE7KGYOwINper/O2DdFujkC3sS42W8AgDDVriPNpqGPngG19laNZP1u3fhM+/qkvzIk1GtzzHZUKxdra2ynWKPJ2AFBLPN5KYmwUxbUsq4GgfuGX+BhVcW2zpqKCGJdIZQrVqrxdvSAa3uJwUfdO5gOn2GejIdrT4oOTOo66/vW6OL8M1HFq8hz5OB++/Ro5VhfzCmULgFdm/f/VZswYY8wScMlFzIi4q1mdY9/ZgTOX+nDGXDCzx+rwucHlbo4x52UxE/gxANtm/X9rMzaHlNI9KaW9KaW9vX39izicMZeW2WN19Zq1y90cY87LYt6Bfw/AFRGxEzMT908B+JmL0qqWkW+DocUm8ek6f35wcIBiIyPnKFarc16eUonfC6rjAkCtysfu6urm2CqOtXdwlarUEMJX8OVN8gVky3IYQp5QXmXMN2Z+l5YqfH2nG/wOdOzcCMXKXbqPi2VRcSwpvYdjNfEeuz5ZpdjkuQmKtbXzO/l6xv00OjFKsULw57u71lAsiX02xHtjNa5afTcNaJFXvQNX/ah2qd53q2Ord+D6HgEa4owaLb5rz+KCJ/CUUi0iPgLg7zCTdewLKaVnLnR/xhhjFsai0smmlB4E8OBFaosxxpgFYCemMcbkFE/gxhiTU3JUkUcIkxlbKiNPEmLTY499j2Jnz5yi2O5d7B7dtm0rxSrC1FHLWOh/6jQvqXz56AsUU8LHFnHsDRs3UiwK4vdzYpNItgTJn0/yOvzgiZj1Rh3DY3PFu2qVBcIzp1n0fvUYj6Fie5c8TvcqXu1SKfA4ErompmvcnkaVxfXxERYhO8rCWFTQ4tnINIuy09PcoF07r6DY7sv53ulQRich3EkxL+OmV+J8QxrRRGghhqEWyBIxC9KM2LpgqfdpjDEml3gCN8aYnOIJ3BhjcooncGOMySk5EjEZJVbOxIVjKrGY+NIRFg1ve8fNFLvqyl0Uaysvrus2rucsptu2cS6w4ydPUGzwHDtDh8+xS7BnzQaKqfSnKtlc1rZaRRLZH/Uuc8Po2Bi+891H5sWEIxHc7xNT3KGTdRY7AaDcxvFiQ1wj0aGTiQXLuhDuutpYNOwQLt32is7uWS9MU2xsjAXUffufoNipM69RbNfOnRTr7+c0Gx2dnRRLwl0JaEekyuQZom8vdtralOGkTMptukgnpp/AjTEmp3gCN8aYnOIJ3BhjcooncGOMySnLLmLKNJLixb42N2nxoSFSvRZUSTVRX6zeYHGmJMqstSo+ZJZUE2XI+tdyOs6OjjaKKWHz1Bl2dtaFmNbbx47NUpmFOEC7xJSGpBxmeU8xW683MDQ6Nw1rEnZIJZiX2rg/O4VoCADFAsfbwNd8UtTHq4nnr5HxMYpNjHGsEixYdidd9q0oml6ucBrcydFJir3wCpUIwNHjPH57VvPY37aVHcfr+vtkG3vWsqO1VBCl6cQ80KrrUon9Oj2t3p8qlabTybZ+7/gJ3BhjcooncGOMySmewI0xJqcs6h14RLwEYARAHUAtpbT3YjTKGGPM+bkYIuZtKSWXmzfGmCVm2VehtIqyp0Zoy2m5xOqzSg1cqbDaf1Ks8GjsuUocm3dYEPm3s3IDt0p7G7exf21vS589e3qIYmNl3t+aPrb1A1lquFDdxSnm3UrfSAkT03PHV1mmTxD9UeeVTAkcA4AoiqK4otunq7zCoyqas6qTi2CPDI9TbHiaCx1PZVi428QYXNXGjSwWebux2hRvJ+zsU2c4PcTQEKcu6OoWRaABbNq0mWKX7+QUGN1tvNKmIs5P5X4XdciRwHONsvADWXnHebus1BaKxb4DTwC+HhGPRcRdi9yXMcaYBbDYJ/CbU0rHImI9gIci4tmU0rdmb9Cc2O8CgC1bL1vk4Yy5dMweq+1dq5e5Ncacn0U9gaeUjjX/PQXgKwDeLLa5J6W0N6W0t7ePM44Zs1KYPVbb2vWf6sasJC54Ao+IrohY9fr3AP4ZgKcvVsOMMca8MYt5hbIBwFeaIl0JwP9IKf3Phe5EvdhXYqASi7LygZ8bGuRgg+31wyKv9mvCXl8qKSs9H2JIHHdqkkUcQOfaVgWDu7q5GO7oqCgyO8Ei1+QYb/fKEAubddE3ANDTx/nE9e98JVnmW8ZspISJqbl9OlVtTaRuF0V7s3QpJQCrYrwqNibyk7d38A4rZRba6lXebnKKhU0AqInFAkm0p01Y1/UjorrHRLFtcYyRcT5nADh36CDFzgzw4rhV7WzZ37qFLftrhTW/TaQPUOO8UdP3U01MVyodgqpdkMUFT+AppSMArrvQzxtjjFkcdmIaY0xO8QRujDE5xRO4McbklBXpxNSOJX6x/8Lh5+Tnf/e3f5tiN1zPr+sbwjF3+uRxiv3d179BsdERFgjPDrKIWanoHMu9PSySlEt8OTZv5ULHddE/zx8+xG0c5IK5oyOcG3rgrBB9Abzl5h+lWEkIOQUxjPKdDXxmDE7Pc9RFXeRzFu7Fhsj1nklFODlF/vlGgYUxMVxQFQ7LthKLqt0iz/z4NAvhAFADH1ukmsdUjYMVke+8KNyLSTxLVoW4XhN50QG98OHE2VMUe22K74nDR1+m2Lp1vOR58+ZtFOvuXkWx9gr3NwAkIfJWkxAxRYHmLPwEbowxOcUTuDHG5BRP4MYYk1M8gRtjTE5ZASKmECyFuDghXIX/5++/Kff4jQe/SrETx16h2A9f9UMU6+/loqlHXzpKsa3bWNC4VhRhXb1aJ0WqiLSWyt5ZrXFftAkX53XXvoliReFUnRbO0IYQmgCgOsGut4YQWEptLNqURCVcJTTp1JtC2FMuPwDq4zqF78Jk1QSglpEWdDbKxTopnLIlpTgCqIumlgrT3B6xXbnMwZK6pVWaWOFy7BbFmAGgJh7zREZYVMVxanU+l4JyIQubYl0IlvVixnUUup9yS0fwOdZEntjh11jYP3r8JYpVxNjv7OyUTVQOXZXKtpxRZFzhJ3BjjMkpnsCNMSaneAI3xpic4gncGGNyyrKLmEmIQIOnjlGsdxU7AH/8x96jd1r7GIWeeopTlR888AzFPvrRj1JMiQonjnMbj7/2KsU2bFApWYE1azit5dSUTj07n/Xr11OsWGSR7+TAWYpVq9zfWc4vJRcp8bUsih90dHIa3F6RolMJPkmIndWs2qIhxE3hbstKPZxFSglT1bkCnBJHVd1Q5SSuZaRqnZjiepVlISYWhfBXKfF2SaR+jSTqNgrBMTX0OFClUcfrPI6mRR8XhPg8LfqxLFTaVOD9VQsZY1W0sSDuCQS7TVX2ajX2G0K5nRZC//BYhpNSCLqY4s8vpI6un8CNMSaneAI3xpic4gncGGNyynkn8Ij4QkScioinZ8V6I+KhiDjU/JdfbhpjjLmktCJi/imAzwD4s1mxuwF8M6X08Yi4u/n/XznfjgLssRsUNRr/8D/8PsU29vdwbNNmeZyNGzdSTAmR3/3uIxSbnmah4cSJExQ7cuQIxc6eZdGwp4fbDQC33norxa666iqKtQmnlqrXp4TNSoVF1ZeEqzRLxJyaYsFn4AzXGVRCYkn0d4dwol122WUUW9vPqTxTm64SXxJClRK0Fkqj0cD45NzzLym1q9Ga83Fi7KQ8TlsbN7Z3Azt6O8QlKgghsSjSxKYCu3nPiVTDE6PDso3bd15JsZEqi9SDg1xjtlJhkbpa5XsshJWyoS6kLjcpt62Lj7eB+6JQFGlrRc3QurKfKlfpFKdsBoDGELvBB47xPKJE+CzOu2VK6VsA5s9MtwO4t/n9vQDuaPmIxhhjLgoX+g58Q0rp9coHJzBTod4YY8wSsmgRM80ses38ozUi7oqIfRGxb2BA/PltzAph9litT7e2Jt+Y5eRCJ/CTEbEJAJr/cu2iJimle1JKe1NKe/v6+N2mMSuF2WO1qLJFGrPCuFAn5gMAPgTg481/OX+rIgDMcxmdEyLml/7iLyjWqLKTrZRxk42Ps7uto4NFMOVG+9jH2MWphERV61I79bQD8ODBgxTbu3cvxW644QaKKcFSsX8/u09rNRZsslLeVqss+IyIWqA1oRaVy9xnXV0saB09yqLqZpGW95obuG8AoNjJx5HpShcobCYk1Of3ldjHWlEjdLU4z4nOjFstWNArj/JYbxc5XdU4mOxgoXhapCTuaOc2Fju1UNwpxkdP1yaKbewXqYrF+J8UF2NcbHfiNAu/1TGeLwCgnPgcSzUW4YsN7u9qVaT/LXL/NMB9K1MxT/D+AGD4tZcoNjXI5zg62vpff60sI/xzAI8AuDIiXo2ID2Nm4n5PRBwC8O7m/40xxiwh530CTyn9dMaP3nWR22KMMWYB2IlpjDE5xRO4McbklGVPJzsxwYJNmxAIq0J8yBIIVe1FdRzFGeE0vPPOOyl2+PBhij311FMUU6IhAFxzzTUUUyLoyZMscqh9KqF10yZ2pJ49y7X+svrx6adZBB0dFekzh1m0USlVlbi8aTOLYW+/6WaKXXXNtbKNxYJwzAlRVeiab0xKQG3umFvTuYo26xHi5LHjL1NsIkNwnxJuyjjBwu7OPhYs12/bQrFnX3uNYqnBfdQ5xvfDmi4W6QDgqVeepFj3RnYbdlfYffvi8wcoVu9iJ3HPFVzTtXvzboqNHWXxHwCKwkW6OvFYHR9lEXR8hBfRtZW7KTY8ya7fjp51FOvr0OlgR4ULVJR/RSjHb0Y6ZD+BG2NMTvEEbowxOcUTuDHG5BRP4MYYk1M8gRtjTE5Z8lUo80XXaWHXVqsi1HZZSwvUKhRV9LdU4tNXtnKVs3rPnj0UUznCBwY47zKgV2lsFRby97yHCzf39fVRrLe3VxyDj/vgg39LsS9+8YuyjYODvGJFrZR58klepaDaqPq7UuHVM6rY9LqNvFoFAN717vdRTBXSbWTnW9OkhEJ97pjb2M0rE04O8gqG6iruo9IqXsECAAWRS71W5X7ffuPVFBsUKxOm1wqLfHC/F1bzipMhsZoIAEYmecVKY5xXc0xN8oqaNeI4r4iVTGOn+T7ZLnLpb76SV6sAwNABts2PHePVPIMnOTY8xseui9QF5yb4unas5VUoq7ZxDABq47xSZnKCbfNq/GbhJ3BjjMkpnsCNMSaneAI3xpic4gncGGNyyrJb6WtCnFRil0IJgVmofV59NQtDzz77LMU++9nPUuyjH/0oxVYJoSpLxHz5ZbZb79+/n2JdXVw8VhVPVnnDt2/fSTGVc/xTn/qUbOPzzz9Pseuvv55ie/b8MMVSg6/N8ePHKTYh7PUDZ05TbDCjmlNNVM4plMX4EeLrG1EqFtG7eu717O/m6zt0llMd9LazpbxS1sevVVn4W385FxHetWkbxZ55mUXzHiEK10QR4fUbWSAs9LNICwBjJX7OK6zi4wye5uLf29ezMD/exu0ZrLM1/+wgj4PCJl5QAABb97yVYsde5Xt5coLHW7nI1yaJdAzFBs9VU0MsYp+GFoNrYqwXity3GTXGJX4CN8aYnOIJ3BhjcooncGOMySmtlFT7QkScioinZ8V+KyKORcT3m1/spjDGGHNJaUUt/FMAnwHwZ/Pin0wp/cFiG6CcfZUKu7cKa9ghOTUl3JkAJqdY2FI5tA8fPkSxbuG2U7nE77//foopUbW9XedYVgWDlbCpHKRK2FTu1TNnWECdnubjZrVRuU1VTu/hodZyjHeLYr/tQnTr7uTtqtM6n/vX/prrad/+fq4CmFTi5TegrVzE9o1z3a3v/7F30nZHj+yg2MgkOw2nJlm4A4DaFI/LHZtZqFOicOrnfO/nhGA5Ns7t2drP+cVrSeecHh1jl2Nq5/zm3YnzfBcbrMhtWMPFk8dOsWA5eozHWnVKt7FrA4ulm6++hWKN6jmKnXrtBYqNjwohUpzL6i7h8IYeq0nMttVx3udCxup5n8BTSt8CwMsejDHGLCuLeQf+kYjY33zFwr96m0TEXRGxLyL2DYhqN8asFGaP1SmR/8OYlcaFTuCfA3A5gOsBHAfwh1kbppTuSSntTSnt7evvv8DDGXPpmT1WK+38Z74xK40LmsBTSidTSvWUUgPAHwN488VtljHGmPNxQU7MiNiUUnrdVncnAK5+m8F8KaZbuBevE26/x777LYr1ipSlWUxOshAzMsJCxerVayhWF9Yo5SocG2M3mfosoEW+Rx99lGKHDrHQ+ta3sutMCai1Gh9jeJhTWqp0uTOfZ4HtlCiyPCr2qdLyKqaE4HzwIBfC/V//+2H5+d1XXkexH/sX76dYeYFP1MVIWF2cO2bediP305uv5sLCI+N8TtWkn5WqNXHdxvn1zcQk73PnNB97fIrH26goYFwWbtVBcR0BoH0nC80T4rqlHv4L+9gJvk8Ovchi/Z61LKq+fFpIbw2darXezvNI9/YbKXbL5TsodvYVFjGfe/wxip068RzFuoIFfEzxPAAAk3Vue4h5oFTm7SZrWgQ/7wQeEX8O4FYA/RHxKoDfBHBrRFyPmfn4JQC/cL79GGOMubicdwJPKfGaLODzl6AtxhhjFoCdmMYYk1M8gRtjTE5Z9nSylQo7utav45py54TAMjyixYLVa1iIXNPDMSUkKrHztttuo9iASG/6ne88QrEsEVPV7VRC5KlTnK7y4YdZ0FPuzM4OjtVEe5TTFNBuUdVnPatZQFJOtoJwlY6P8zVUbs+GqP8IAOeEC3RKXMOFipiNWg2jZ+fu+9UXWavfuoVT9m7ZtIFipU5dE7Mh6lUOC7/EkDjPvl4W8ccm+JqNTwh35ij3+8go3yMAcOXlu/jzQrCfFONoXQff32XhoP6Rt7ydYmfHebuXTrCTEgCmC+wmrk/wOICoYbn5TXwN172Ja9HWBlnAP3uQFx68+PT3ZBvPvMDpmQtt3I+FkhjrU1rE9BO4McbkFE/gxhiTUzyBG2NMTvEEbowxOWXZRcyCchCKOoGFAje1LgQ1ABgcZAfXyDCLH8qN9t73cmrzNas5V9djjz1BsW5RM1G5GRdCWxu74BTqOFNChAwhnrZ3stgJAIVpFk5UQrIeIRrv2r2bYqou6blzQxR74Qg744bPCccbtOCpUhS3Xj11hmKhiJ55IvDIANd8PC7GYP9GPv6aor7VulZxbUqs4XFUDL6Wq4Quu0aMwVRorU7mwQNcQxIA1olFBZ2d7EodF8LodTvYLfqje9khOSEcqePi1rlim14UcHKABdTXTvA8cOLFVyj2sqh/OSlE544eTlnbc817KXb9lW+TbdzyIte83f+dByl2+sSL4tPaJesncGOMySmewI0xJqd4AjfGmJziCdwYY3LK0oqYCYh5omW5xO68TlETEeDtCsWM2nGN1mrKqd9eb7r2Woo9+9wRinV3scgxKaq4KOcioJ2Yalvl5FSCpRIIlesy1Vs7BqBTvUaB+3ZkjGsuqjTByml67lV2bJbKZfFZ2URcccUVFGsX9RoXSrlYxKbeueJsiHqiZ0+yU/bJ/Ycp9sTTnIoUADZs2UaxW370HRTbso6F4slBFnCLJaFsChFTjZfLNuvCWh3tfD0qbTx+V7eJ+3YVH7ta5+OMCAfpRJ3H2sFDL8k2Dk5xTc0bd7H4Orqez/vF4yxOHzzKgu6TR/i6jlRYhO5freYvYM8GFnT3voMdn0888hDFhod4nAF+AjfGmNziCdwYY3KKJ3BjjMkp553AI2JbRDwcEQci4pmI+MVmvDciHoqIQ81/MyvTG2OMufi0ImLWAPxySunxiFgF4LGIeAjAvwbwzZTSxyPibgB3A/iVhTagrY0FJyViFoSTrZpRJ65cYtFFOc+UWFarsaCnRL6NmzZS7NBhLVQplGCpHIRKsJwWDkmV3jOlgdYak6X5CuGwUOTf+Wu3bKLYoUPcF6dPs9CkUtYqgbfSzulCAeDHf+J2sS0LedUFWjEnxsew/4m5aUHTwFHabk0fC2WPPcMC2LMZ4ttNt72LYv/9vv9GsR9/180UW9vOJ9XeweJxqcz308QkC6Dr+rguJQA0KuzUHRQCtyLEeKmK58Yo8/U9fPRVin3yE5+Uxzlzil2Xb3kr99m//MkPUmz9Rr6GXTVekLC5xjfKM0N8HzcK2n196mUeP1dcxqmHd125h2LPP8Vpa4EWnsBTSsdTSo83vx8BcBDAFgC3A7i3udm9AO44376MMcZcPBb0DjwidgC4AcCjADbMqkx/AgD/KjHGGHPJaHkCj4huAF8G8EsppTmZVdLMAl/5R2pE3BUR+yJin6piY8xKYfZYVYnAjFlptDSBR0QZM5P3fSml+5vhkxGxqfnzTQDkSvOU0j0ppb0ppb19ff0Xo83GXBJmj9WK0EeMWWm0sgolAHwewMGU0idm/egBAB9qfv8hAF+9+M0zxhiTRSurUG4C8EEAT0XE95uxXwPwcQBfiogPAzgK4F+1csD571naxYqBo6+w+lzpYCV9bZd+oj99kq2xamVCp7DD79x1OcVePcbFTEdGOL/4li1slX3xRbbhA9pWrlahqBUZygat9pcavHpGHSMLtc+SSH1w7Ngxiql2qxUnaoWPOm5BnDMANMSLu4Xm/lZU6w2cHpq7UuPZMq+iKZ7ilT4vHz9OsXe861Z5nF/7jV+n2Kc/81mK/c1fP0Cxq7ZwUeNyG1+frlWrKab6vXdNr2zjul5RpFlcD5W7viCKNo/WxcqqEo+Xz/3Rn1DswLNPyTZWynzsrzzwlxTbeiWnyrj2ih+iWEeFV8WsTtzuzd3clpo4FwAYE6kB0jSv5tm+hXOtZ3HeCTyl9G1kLzTjNVDGGGOWBDsxjTEmp3gCN8aYnOIJ3BhjcsrSFzWepzBVOthKX6mwYHnTzZwj+ed+/ufkIT79n/4jxR77HltRk8gxfuYMi1KrRG7rgQEWTRpKUctAiYkdHSy0btjAAtLJkyyqyrzhwj8eIcTOBah+8jgix3ir+c4LIr94IwnJJfhaAcDgkCr2qiSbhUmbbZUKtuyYW5i5Ds5dXq1O8me7WNnatI0FbgBI4nps28zFc7/x1S9TbOQEpx/qVPeTGFeqjyoiBQUAdHfy+XSKRQVtQkhsb+NjJ5Gv/fQE9+0zBw9Q7N3v1rLbdddfR7E//q8sgj7yrb+l2K6NnNO7rVPMDSd4ccSTh56nWLlL9TewYTUfpz7B91OHyLWehZ/AjTEmp3gCN8aYnOIJ3BhjcooncGOMySlLL2LOE5NKZRY0fvN3PkaxhnAVbt3KYg8AXLXnaor9l8/+Z4p9+a++RLFzw1ygVwk+/evYBVoXDjPlWAN0Tm/lZOvuZgHp6FHOK6zERSUQQgiJWSKmdEQKcTIKQmAM8WwgmjNd5T5Loo2X7d4l23i5cNElKWLq4tJZJCTUMLdP60KkbhOCexcbHzE8yvm3AeDkKXZ3njk7SLFXT7C4nmrsbG2viFzoVeF2FW2plPVY7aqwuFkUjtwOkbO9vZ37pyGKkb98moV5CDH7jjvvlG18+9vfTrFXhKP7Kw/8NcWeeHI7xeqTfH8OnmT39fQAu5BLdV70AADjNZ5bjgy+QrHOCs8DWfgJ3BhjcooncGOMySmewI0xJqd4AjfGmJyyDCLmfGGCf4esW88FgxUTU7pqivr8r/wqp+3ctm0bxfrWslvqO9/+NsWGhoYotnv3brEdC1KAFh1VYWIlWLaaYlYKjiqdbFaKWSFiqnZP13m72jRvVxYi7c7LWZx857vYbXfHB35SNvHaa36EYsoRu4AsugBmilufGZorHFZr7LosiT5OojD2E/uflse59jpu/xP7OWWqKgQ8XWLBcrrK4uLx41wJa3JKOEgzBPey0qjVdm0sdpaFMFpPLCiPTnIR4d5+diH393EKXQAYGWZHrio8fnaQReOvf/1Bik2O8r04MMAi5JgQ60vCDQsARSHKrt3ABZXXb2ht/gP8BG6MMbnFE7gxxuQUT+DGGJNTWqmJuS0iHo6IAxHxTET8YjP+WxFxLCK+3/x636VvrjHGmNdpRcSsAfjllNLjEbEKwGMR8VDzZ59MKf3B4prAL/arQgSSskmGMqU+XmpjYeGmW26l2Ffv/yuKnTvHDqzTp09x7AzHikWdBlWhBMKJCRZ3WhUnk0jBOq06R4hKWfts7+ii2EYhGl/xQ1dS7JZbbqHYbbfdRrFdQtgsFLU7rSpS5iYlYi6wUGaKhHrM7ZcQbRgdZ4flxCiLXSdOs7r7C2sAAAomSURBVJMSAD716c9Q7OhhFq5HhSh8+BgLcurc1biqihTAUef6jABQFM95Ie7HEKlRU7DTVt61QjDv6OL2DAzofqwIgXz4HAubU1PcnpdeYsdmiPukKm6TJJymWUNNpdvtqrDTenxMzX+aVmpiHgdwvPn9SEQcBKCTGxtjjFkyFvQOPCJ2ALgBwOvVET4SEfsj4gsRwdnljTHGXDJansAjohvAlwH8UkppGMDnAFwO4HrMPKH/Ycbn7oqIfRGxb2CA16Mas1KYPVZr09pjYMxKoqUJPCLKmJm870sp3Q8AKaWTKaV6SqkB4I8BvFl9NqV0T0ppb0ppb18fZ/AzZqUwe6yWhCnFmJXGed+Bx4yS9XkAB1NKn5gV39R8Pw4AdwLQVrPzoMofqlSkSQkpGXlQG0mkVhXSSU9vL8VeOsrpHe+44w6Kfe1rnJZyTc8aio2MqJqNWiAcG2PxS6WdVZ9dLep2bty2g2Lbtu+k2NatWtK47LLLKLZ79xW8z53sQFUu164uFkBVncxajYWmKdEPAFAMMdEu0HWpKJVK6O2bPz5YFJ4Qjr0pUROzoNLrAhgaZEdv37r1FFvTy469mhAsG4n7qVZlMbAu+lilnQWAhhCKlTA6NcXHbqh7VIjmBXF/Dwl35T985x9kG5UY/syBgxQTzca06MeiuNYNcQ2VGFzPcIhjmo/ziphvihWdjlbRyiqUmwB8EMBTEfH9ZuzXAPx0RFyPGdH1JQC/0PJRjTHGLJpWVqF8G/qZhhMIGGOMWTLsxDTGmJziCdwYY3LKMqSTnQ+/nWnINzbCXScciYD+raRqavb08NL1DRvZVXjyJNfru+nmmyl2771/QrHpjOVot9/Otf2279hBsWqVP98pxMDLhGi4ffdVFFu3cRPvr5PdZABQLotaiEUeMlXl/hPi5LioM6gudUGItIWCHqriMBflqSQhoT6vjqYSXEuifmFF1MnMqo26dq1YmSVcgCpFbkG4fGvT7Axt1Lnf60J8U+cH6JqpNVHLdFSI8FNTLKBWq6I94pzVZ7/2N38j2/j0gQMU2/fY4xSLAo/puhiENZVKWYivqSb6UdTGBWYs7fMpiHqy7an1Jax+AjfGmJziCdwYY3KKJ3BjjMkpnsCNMSanLKmImaCcl0KcyUzIOP+jGdsJYSxJYYwFjdvvfD/F/s3P/az4LP/uO3L4EMX6+nX6gA/+PPuerr3uRoopYalQbO33rkrLq8RFJWgBwNS0SDkaLEAlYadVbtGiSG+b1LUWochwMqqMwkmlx12gOzMQiHntLZeFG7godlznmBKEAchzVWO1otISi+3axB0daKeYEiHV2JhpUGsCal8/O5ur4jjq+mhRVdWNZZEWAE6IhQY7drDreGSMBcJxkbJZXZiWhc2MflR9puaRQoGv6/iwTqPrJ3BjjMkpnsCNMSaneAI3xpic4gncGGNyiidwY4zJKSvASr8Y9CqUrMUp81G5in94z9UU+3e/+TsUu++++yi2Zm0fxT7wgQ/IY+/czUV/x0QeYZlUYFpbdS+UhazwWMx28rOtLg9p9aJice35f4dDIKW5qwZSQ6y2UcV9xfGzbOpydUqJVyuoVT0q5YD6bFGsdCgLa75K2wDo3N9ypZfKqy3ytdeE1VwtsimLdnes6pFt3HIZpzRQ6QcmRHFotVJGXa8Qq7+SGJdZ11oVONd51TmFwLGjL8p9+gncGGNyiidwY4zJKeedwCOiPSL+MSKejIhnIuK3m/GdEfFoRByOiL+ICP4bxhhjzCWjlSfwKQDvTCldh5kK9O+NiLcC+D0An0wp7QYwCODDl66Zxhhj5tNKSbUE4PVEv+XmVwLwTgA/04zfC+C3AHzu4jdxiRHC0E8Ie/0tt72TYuPjbPPtz7DSz7dpA0CGkVl89iKodOYNSY2E6cm5ApNMDyAegZT4lilsiTzhIYRIlXKgoXLkC0G6IITEcgfHUlGLmJUWUzcoZVOJfKpodVUUrW4Im7r6LACMi7QPSiCcrImFAup+EikSktifss23temXEVk54eeTlZ9f0dKViYhis6DxKQAPAXgBwFBK6fXefBWALm1ujDHmktDSBJ5SqqeUrgewFcCbAXCplwwi4q6I2BcR+84OnLnAZhpz6Zk9VtUToTErjQWtQkkpDQF4GMDbAPRExOt/E2wFcCzjM/eklPamlPb29unXCcasBGaP1XLGn8HGrCRaWYWyLiJ6mt93AHgPgIOYmchfd6l8CMBXL1UjjTHGMK28Vd8E4N6YUd0KAL6UUvpaRBwA8MWI+F0ATwD4/CVs55KRhAg0IZxaHV0sNHSvXk2xmsjJDWjXmsoDbJYPznPO10cV40VwrFKpyGMo92O9zrFyG4uOssgyRNFeMX5rKg95httViaVqrCoxUBUeL1eEW7TMf/Go/UlXKHRfVIVgWWgI16XYZ03EiiLvfUOIqln9mBWfj8oRnkUrq1D2A7hBxI9g5n24McaYZcBOTGOMySmewI0xJqd4AjfGmJwSrb5YvygHizgN4Gjzv/0AflAWhvtcVibnO5ftKaV16gceq7ngn9K5yLG6pBP4nANH7Esp7V2Wg19kfC4rk4t1Lu6TlYnPxa9QjDEmt3gCN8aYnLKcE/g9y3jsi43PZWVysc7FfbIy+Sd/Lsv2DtwYY8zi8CsUY4zJKUs+gUfEeyPiuWYptruX+viLJSK+EBGnIuLpWbHeiHgoIg41/127nG1shYjYFhEPR8SBZqm8X2zGc3cuwKUp/eexujLwWH0DUkpL9gWgiJliELsAtAF4EsCepWzDRTiHdwC4EcDTs2K/D+Du5vd3A/i95W5nC+exCcCNze9XAXgewJ48nkuzrQGgu/l9GcCjAN4K4EsAfqoZ/yMA/7bF/XmsrpAvj9U32NcSN/xtAP5u1v9/FcCvLneHXsB57Jh3UzwHYFPz+00AnlvuNl7AOX0VM6mCfxDOpRPA4wDeghlzRKkZnzP+zrMPj9UV+uWx+v+/lvoVyhYAr8z6/w9KKbYNKaXjze9PANiwnI1ZKBGxAzMZJx9Fjs/lIpf+81hdgXiszsUi5kUmzfz6zM3SnojoBvBlAL+UUhqe/bO8nUtaROm/f4rk7fp6rDJLPYEfA7Bt1v8zS7HljJMRsQkAmv+eWub2tERElDFzQ9yXUrq/Gc7lucwmXUDpP4HH6grCY1Wz1BP49wBc0VRb2wD8FIAHlrgNl4IHMFNWDshJebmYKXfyeQAHU0qfmPWj3J0LcElK/3msrhA8Vt+AZXhp/z7MqMgvAPj15RYRLqD9fw7gOIAqZt5TfRhAH4BvAjgE4BsAepe7nS2cx82Y+ZNzP4DvN7/el8dzaZ7PmzBT2m8/gKcB/PtmfBeAfwRwGMBfAqgsYJ8eqyvgy2M1+8tOTGOMySkWMY0xJqd4AjfGmJziCdwYY3KKJ3BjjMkpnsCNMSaneAI3xpic4gncGGNyiidwY4zJKf8XPczBzPY+vAkAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAADTCAYAAACY5kjLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO2daYwl2Vmm3y8i7pJ3ybUqs7bsru52L24bd9su2g22jI0xIP/AoBmNbGaQR7LUCMkSlvhBw0gzjIQ0ZsTyY5AYNbKFkRizCCwbj8fQg8xgkKft3txdvVQv1bVlLblvd1/O/MhrT+Z931t1a8vMgO+RUlX1VdyIEydOnBsZ73m/z0IIcBzHcdJHtNcNcBzHcW4Mn8Adx3FSik/gjuM4KcUncMdxnJTiE7jjOE5K8QnccRwnpfgE7jiOk1J8Ak8RZvb3ZlY3s83ez6m9bpPjKHys7g4+gaePz4QQSr2f+/e6MY5zFXys3mZ8Anccx0kpPoGnj/9iZotm9k9m9qG9bozjXAUfq7cZ81wo6cHM3gfgZQBNAJ8A8PsAHg4hvLmnDXOcPnys7g4+gacYM/sGgP8ZQvhve90Wx7kaPlZvD/4KJd0EALbXjXCcIfCxehvwCTwlmNm4mf2UmeXNLDGzfwvggwC+sddtc5zt+FjdPZK9boAzNBkAvwngAQAdAK8C+NkQwmt72irHYXys7hL+DtxxHCel+CsUx3GclOITuOM4TkrxCdxxHCel+ATuOI6TUm5qAjeznzazU2b2hpk9fqsa5TiO41ybG16FYmYxgNcAfBTABQDfBfDJEMLLgz5TLIyEybFy/35ou263qw5IoSTW3z9qnx21zyBi4jtN9VGn26FYJsmI/en+jSI+TqfL2wbx+a7YDkNeR/1R1Q9AV+wzElYM1d9mfH6ZhFetJiKmUMcA9LUZdky/dW5uMYRwUP1fEkUh2ze+1F7l8cV2qj8AIE5itbU4Dl+jKOLPqvaorotEe3RbgDjmeLfD47/VbovPinGQ4ftEnUun3aJYUPcx9PA3cY9lslmKqfmm2WjK4/Sj+iYaMC+p+6ler1NMzVWNRlOO1ZtZB/4IgDdCCKcBwMz+FMDHsZX/QDI5VsZnP/WvdsTyI3wxq1XReeICj4+X5HGyYiLdrFQp1mlzzOIcxZotHpibGxsUm5meoVhXTPQAUCiOUKxSafCxxY1ba/LAbjdqFIvEnVtv8CCqNfVgbbR4n/kcXwf1RTqSKVJs+uA0xQ4ePCCP3U+U4RsPANqqL9p8PuqL8N/90uNnBx0vG0d42/j4jljHeB8NMcnIx4JcXh5nbHyCtxUTT6PBY2N0dJRirSZvl4j95fPcnokJbgsAlEfLFNvcXKfYwtICxUolvkePHDky1HZL81co1qpWZBvbotOTHN9jR2fvpFi1yuP8wlkeGl0xhspjYxQrif4CgLq4hi+f4jTpaq469cYZOVZv5hXKUQDnt/37Qi/mOI7j7AK33YlpZo8BeAwAJkb1E7Pj7Ae2j9WMeGp1nP3GzYzSOQCz2/59rBfbQQjhiRDCiRDCiWKBf6VxnP3C9rGaqJf9jrPPuJkn8O8CuNfM7sLWxP0JAD9/tQ+EAHTDzhujXud3iBnxTrUlXnI1O/p9WMjxu6q8eB8WEn5XJd+lRdxNRSGG5LP8/lxLLkBZvPOLY3533xDv5wrG7SlNTlIsk3A/bNa4vyt1fjcHQKpfI0V+71qvq/fvfOblMr8XHxnh66Le987eeYduY+CxUqtuUqzZYLHo6hiMJnHuTyXIJTkeG0FcW0CLs20hBh48yFrroUOHKLa4KN4bN/ncp6d5vMwe029A8+J85i9zG6Mu30/lsXGKTY3zGMqIe6d4h3hfvcHXFgBqQnRsiZvv9NnzFDt77hzFNip8HCX8jpZWKXb08GHZxkNCIxvLFijWqQ8noAI3MYGHENpm9hkAfwMgBvCFEMJLN7o/x3Ec5/q4qXfgIYSvA/j6LWqL4ziOcx24UuM4jpNSfAJ3HMdJKbta0CGKDLn8TrHCYhZxWlUWsSIhTKLAghwAbAhxcyJh40K7rexbwi0ohKY45u3iRHwfdrSMWamwsBSrlQ/CCKRWuAXx0Vqdhab19TWK1Ru6Hw8cnOLYJAtVwtMknXFKnG63WLDJCdEsK0QuQLvtmi3Rt139+UEEBHT6REvlpMsXWYTKCJNMd4ATsyMcjcrdVyyyAJwVQvrEJJtxWm2+n0bKLKKXJlhwBIB8jvuuKkTDbFHss8TjZVyYX5rCkKUcx8sbLNICwJWFpaFiZy9cpFhbmOVGJ7kvgjD0LZ69QLHNdTbiAEAu5nFRzHKsnhl+rPoTuOM4TkrxCdxxHCel+ATuOI6TUnwCdxzHSSm7KmKaGXLZnY60dlekoMzw90pSVqqYFt+KEQtLScSCT7PLYmfUZcdcTmQoVO62msgIGA9Il9qssAik0lpmhdNvpCSEM5GBMYwIgURYTS3WLsVWiwW2jXXOQqfchHnlsBSOz1ZjhWKH72DXZUuInYBOs5of4euVFU7cqxEC0OzsHJtJlvs4FjFT6UQHWHKViFkoCHee2K5S4fHb6Yh+EtdnYZmV55FJfT9FIs3sYpXPMQK3u1Lh+3bikMjgV+TtTr3yCsW+d/JV2cbFZR5HQTiocwUWVQtCDC6UWDSeEW7YpcvzFFtf5YUCAPDqG6cpNj3Gwq9KuzwIfwJ3HMdJKT6BO47jpBSfwB3HcVKKT+CO4zgpZVdFzBACWn1uprooAaWEu0i4sroN3fxSmR2EaCoBisWdDIQY2BZOzAwfW4l0naBLqgkDqiylNSJKr61ssli6uMwOs3fcf5xiZSGAZsW5AJDil3LHqXqPOVECTZXerNU4bWd1kwU2VcsT0LUZg3C0WqTTuQ7EuLZhTgizsRirXXFtVb3IHxyof5+qzqLo481N7rtGjYXNEeGGrIGF3hdPL+oWjrC7sy1SMXfFWA/LLOg17DLFMk0Wx5/9ztMUe0s4HwFgfJIFxqkDXMIvFoLl5ib3WVOUdayLVMwmXJOlKV0mcGmNU89a4H1OCRfoIPwJ3HEcJ6X4BO44jpNSfAJ3HMdJKTf1DtzMzgDYANAB0A4hnLgVjXIcx3Guza0QMT8cQtDqh+M4jnPb2HUrfUz2VlZh8zleKaEq2sdZVusBYCTDNthuzAp5ts37NFEkNwiLuxnb1HOijY2atoDLAsgi33QkViQsr7Ca/Wd//bcUe+jNt1Hsx3743RRTq34AnfO8C+7HdodXWDREfmeVLz2b535QtvFaVRewzolc1bUK52NuilzOVyOOY4yO71wNoFa3qJU6kdhOFW8GgG6Hr/mGWIXTn0cf0PnANza4n1Qh63lhhW8W9BvV2R96kGLtHFvA1zd4XCZivHz35Ou8vyUuLFxZXqaYCXs8ABSKvCqmWOLiyW2x6mfqgLCzi1VPC0v8nLpR57HWFWMCADIitUUywve3qboCA7jZd+ABwN+a2TNm9thN7stxHMe5Dm72CfwDIYQ5M5sG8KSZvRpC+IftG/Qm9scAYHKMvxEdZ7+wfaxmRQInx9lv3NQTeAhhrvfnPIAvA3hEbPNECOFECOFESZhSHGe/sH2sJuLVlePsN254AjezopmVv/93AD8J4OStapjjOI5zdW7mFcoMgC/3ckEnAP5HCOEbV/2EGVlZx/MsxMjc1qr+sBAcAaAmLMbKx90RQhtELt7ixCTFGhtsEVaFWeMBv4pncnyOqv5xRYh3Gxt8fspq/p3nXqbYsSOHKfbh971XtlGkDtcCW4bP8dRbbHn+9rPPc3sOz1Ds0ff+EMWmRYFlAEiEAJsVguWgosiDiKIIub7fGJUQakKwlG3KsIAFAEGIausVtpW3IxYDmx0W0C6scF5sE9kcah2+74oHud8BIDd5N8UiMd6SGvfP5sqbFGsssBhYEvkHMqIQ+UbQBYNjJSaDx2VbLEg4MnuUYlN5IWyK6zq/xPnA6y1OqQEAR4/yWD80JkT8AZ9X3PAEHkI4DeChG/284ziOc3O4E9NxHCel+ATuOI6TUnwCdxzHSSm77MQE+lNPNxos/CnnY36ERZfIBhQMFrmXg1D5NutczNdETu61dRYSc1nlUhRFicV2ABAJFag+ZF+MivX0b7/vPoqtr7Iz7rvP8UKhRx56h2xjIoS/EPh83jx3iWJ//Bdfodj5i5wH+u7ZWYrdfxcXNS4XtQgougdVIfzGIj/51QgIaPU5QjOigHEQ6nosxlB/bvHvk4h80uM5Ft8qDR6rZ88L92KDO2Qsx33XFfdIbpzzfgMAiizoJUJ0LDQ4j/W6eEZsCXU8EY7a4hTn+J5b4LEGAGtrLPxmhSP7zgffTrGc6J/1DRbrx8f5/Fptvi4qLzoATJS5H4WZFg1VmHoA/gTuOI6TUnwCdxzHSSk+gTuO46QUn8Adx3FSyq6KmJ12F2srO8WGkQILDUosUi6oAVkboerfJiId56hwfEI46wqiEHBXCDGdrhBPRWpUAKg3eNu2EFqnDnCB1Lo49roomLo0v0Cx4gEWhr71zCuyjYWs+H4Xtr7nX3qL2yPcsLPCBXr/fXfxcUXazZVldhgCQDbHBZ6VY1Klnb0a3W4X1epO19/UFLtBlbCViHG1NKD9QYxWJcxenGMBeH2NxdpSiV3DoStSAKubJK/TCjfFOJCFvktjFLvvh99Psct5PsHq+Tf4GCJNsSqMDQCrYvxPTvL1UsXEX32Vx/94nuelvFiQUCzydpWKcIIDOP0Gu1IPTvDcUr6OnFH+BO44jpNSfAJ3HMdJKT6BO47jpBSfwB3HcVLKroqYMADRTpdZJNKtlstc386Mv2uqA8SCINRNE7FYpEHNCMdeJI6dL7HQoIoArCxxXT8A6AphKaNSzwqnX7nIwkdepOVtKJdgg11eZ+Y5NS4A5LIsJm7W2KG2IOowHhLiq0q9+tCD91Bs5iALcaoOKAA0apx6MxLXsFEfPkUnsDXe+oXPQoH7fXp6mmLKnVmtsdgKAEvivK4sLvF2i7xdLs/tSWIWIivrfH07wqVINukfHIj32RbjN1fi+xaihmX5jvsptnT5CsU2N7kfLOFzBrZE5346XR7rZ99isbQu6tZmx4UAKhTQj3z4gxR77ZReFHDyBR4DOeF2tgEOc4U/gTuO46QUn8Adx3FSik/gjuM4KeWaE7iZfcHM5s3s5LbYpJk9aWav9/4ckMbMcRzHuV0M87b8jwD8PoA/3hZ7HMDfhRA+Z2aP9/79q9faUWQRsn3CUEuktWwIoS0WAqENqByu0nkqgbEjhI8kI8RSkZ709BzXwjt6iIW7sQlO/bp1bI4tL7Foc1nUD1TpL0tC+F3Z5HYXRtk5GJmy5QHTM1zDL7Mq6gJevEixjtjloUn+np8a4xSbbeFe7R83PzhOLNx6kaiFKK2Hg4njGKOjO69dqcRtbTRYHF0U17HdFvVXAVQqfI0WF1j4NiEGZjIs6LWa3Hd10ca2CfFMFVYFEKuwuvWESL3e5vupNPsAxe6IWHiunX+N97fIjlQAaDVY5F1d5XunLly65Qke511xLvc+wOLrBz/IIubFc2dlGxOxGEI5S1st7d5WXPMJPITwDwD6R9THAXyx9/cvAvjZoY/oOI7j3BJu9B34TAjh+5nVL2OrQr3jOI6zi9y0iBm2Fr0O/P3UzB4zs6fN7OnNSnXQZo6z52wfqyohluPsN250Ar9iZocBoPcnvxDuEUJ4IoRwIoRwoiQMKI6zX9g+VrODTC2Os4+40VH6VQCfAvC53p9cAFFgkWGkL1WoSh2bFYXilHMxHlBnUNW/NPFLghI7B6Wr7Oe1NzmF6lvnWcz78Ucflp9X4uRTz79EsYUlFmfe8SDXsDw0ezfFoqeeoVhJpGrdWNZu0db0IYp1WiwwZ4Ro2BGu0kff+0MUGx/jFKSNphb8FCPioUDVmewo1fga9Jt3lduvJfqjKoTJWLlsAQQx4JotPv9EuEtjIfw1RHu6whka2nw/mRBAASCuc7zTYldhEG5giLSsocSxY0c5rXBtjEXv0099U7axUeXfmDY3+d7pdPmeH5/kNMctkcp25hDfDyrN9flzXKsUADZF3U4D98WBA+wCHcQwywi/BODbAO43swtm9mlsTdwfNbPXAfxE79+O4zjOLnLNJ/AQwicH/NdHbnFbHMdxnOvAnZiO4zgpxSdwx3GclLKrUnscRSgUdqZhTYTar9K3KhFTpZgFgADeVjk+u112W3WE4BMJsXNqjMWzF1+7QLFvDWjji6IO38UFrpu4KQSxu+9/J8XyI6pupxDDRErMZIDAtrIiHIGif8bKfOz3vJNF1bffy0KVEpeThMWwjBC2ASArRNlsjlP9tq5DGAW2rnmhsNN5qcbQxganNFbX7OA0u3QBoCWWK7aV4BoJgVGMraJwi3bq3B+NDl/zaEDK3UyD+65RYUFubJztIOt5vj6NRLQnx47lTp4F7iSv60XWNsX5CHFdpYS9dGmOYqUmL3lW9UtVrdVD09oW8+oL36NYEvN5HxCpmAfhT+CO4zgpxSdwx3GclOITuOM4TkrxCdxxHCel+ATuOI6TUna5qLEhTnYesi2UfQMrxZHI553P6xzRbWFFrlVZVe6I7bodUfw4Ybvs1Cir/ejy/r79HCvPALC0witOEmHLbTbqFBMLQbCwwOlo2sJWvbbO9uKRos5ZfkC0Z2yM84lX1vjYJ971dorlC7yCoLLOKzZyOX6uGLQKRfVZQ/RZu3l9yanMIiTxzhUUS6u8KmdTjKtu4LFaret8bxtVsTpGrC5pi7FVb/EKmAMlLghdz/FY3dgQq1/WFmQbm1e4EPDKHK/cyLe4eHL2OBetNjEO1GqkIFJlxOq+A5BtcXHpbI7v5fEp/ny9zatvRvK8EmisLAqZixVqs0eOyDYemmGL/OwxXrEyNS6KQw/An8Adx3FSik/gjuM4KcUncMdxnJTiE7jjOE5K2VUR0wBEfcJER9jZlZU+EUJie1DVFKkXieOIry9ll15aYXEmEvsbL7Jt+PW3BhQ4FSkEGnUW31S6gLV1tjE3ReFaVbR5VQhxYcD3+H3330exw0dmKbZy5QzF8iLPuypWnR8RgpY4Z1U8GAA6Ip92V+SDDwNSGgyi2+2iVtt5PUxU8lV1gNsieO4C54oHgNVVHltJwmOjIwZ1tbZBsWaJ0xrEWRb7CyWObS6el218q85C7eI857O/fOFVio1f4bQP9/3ohyhWPMbCXTLB+cBzk5yTGwAiIVyPijUO5QkWecdGeKy2qiwQX7jAeb6fe+4Fim1UOVc6ANx1Dwu6UeBxrYqED8KfwB3HcVKKT+CO4zgpxSdwx3GclDJMSbUvmNm8mZ3cFvsNM5szs+d7Px+7vc10HMdx+hlGxPwjAL8P4I/74r8XQvjt6zlYCAGtPvejKhSr9CYl0gVlSYQuiqxyLJvx6WeEuzOpswsuJxyAR0Ux0ueFIAUAbSGCxjHvc1QoMVXh/ouFU7VcYmEo6q/UC2CA6ouXTnKR5Qvn2YE3O80uuEZFFL3NcBsPTHHuY5W7uxt0wd1KhcWmnMiNnitw7GqEENDpy6febrMIuygcsMqd2RQ5tQEgI8T5g0K8W15jB21dOG1rQsyziMdgqcj3yMYq57MHgM4Si2pdMVarqyyqtr7HfZHr8HjLfoTHeUn0w+QsF/QGgMsXebFAu84u34mD76XY3e96F8W++3++RrGvfO0bvN1TJylWLmiH+GSZ+9wavGiiXuV2D+KaT+AhhH8AoMuWO47jOHvGzbwD/4yZvdB7xcJflT3M7DEze9rMnl7fHP6bxXF2m+1jtf83RcfZj9zoBP4HAO4B8DCASwB+Z9CGIYQnQggnQggnRkvFGzyc49x+to/VjHhF5jj7jRuawEMIV0IInRBCF8AfAnjk1jbLcRzHuRY35MQ0s8MhhEu9f/4cAH6TP/CzfQHlZGuzYNXp8K+0Ku0sAOTyLBaUxNN/p8PHicWTV0EIYG3hmqy0ROrXAQ7ARBRcjYxjqmjq5cuXKXbs6FGKHZxm19ra6hLFJiZYfAWA06c5jWhosZj8yZ/5KMVksWHRj6oAcKkk0tvqS428OE5DiKADXbsDMAOSaOdBm3UWZnNZvoWShMXjelU7SWdmOJ1oIkT49Q12bOZzSqznc2+JcVkQhZ9z0K+NWkK87UY8LjttvnfKIo3uyivPUeyNSU5TfP/7P0KxQ/c+JNt45YV/pNimEFXDCPf3nQ9+gPd3kV2p/+vLf02xtTWeQ+6aZVF/6+A8Vo+UOVYs6dTJimtO4Gb2JQAfAnDAzC4A+E8APmRmD2Nr+j0D4BeHPqLjOI5zS7jmBB5C+KQIf/42tMVxHMe5DtyJ6TiOk1J8Anccx0kpu5pONoSAdp/AZMoZGIRLUbjJBjkxWy12XeZEfTxhAkUQjs2mWBO8ItKAzi+zaNIWQikAmEg5GglhMyPSzi4J91+twseenmEx5d5776bY5CSn2ASAOGExZUykzJ2ZOUixrHC05ke4HmG7xbVBIyEMxgPEYFUzVOV4Vel/r4YBSPoOeVic54GD3HexuGbdARpqUwiuFy6x8zGJuf0HJll8zuX5+ly5zONlosQi5sTYYdnGyxfYfbu4yQ7YTIbHy2iGFw9UOixcr82xaLh8hVPWlsa1QHjoHSco9tazLL6ePMP7XP36tyjWWOLzy46yADo6zoJ7VoxfAKhW2E3bGmMnsmWGF9z9CdxxHCel+ATuOI6TUnwCdxzHSSk+gTuO46SUXRUxu92AWm2nIy2TZeGuINx1qr5jNwz4/hGpR6sV4c5r8z6VC7Qt6mTWauzK0w5SoZQCMGOhTTk+VX3EbpePc/r0aYodP36cYj/zMx+n2MVLXOsPAEKb21hMWEyriRqAY2NjFBsX6UEj0Q9tca1HRO1MQBs0VW3RLvR1GEwg5VHVajWxX+XOLI0JdymAtXUWn8fKLPxdWbhCsemDLGLeexeL1HOjLBAeO8a1TY/N3inbePKFlyn2T//32xQLIr1tO3A62WqL+7HY5HvMqry/9Vg7Wqfe9SjFKjH34+JZFnQXn3mRYhFYxDx654MU66yzKzqJ9Vh76IF7OZjh8b+4zvfyIPwJ3HEcJ6X4BO44jpNSfAJ3HMdJKT6BO47jpJRdFTFhQNRXu3GkxO48Vd+xXRP19pra5dgSBs0g5C4lTvY7RQHARHtUwv9EtbujXVU5kTJUiZgKVfNT9dncRXb0rQsH3dQUOwwBoCa2ra8uUCwSonFDCFqttujbiJ8hMsKRWm/w/gCg3WRRK6P6J6PrFA4kAKFPTK2JWpdtce5B5EheW2bnLqBT+d5zzz0UW1jiNMCZmG/ftx1nIbIo0hRbzH3Uqul0sqOj7DY9PMmOyPkFrqlZbbHAXYv42JPCxZkTTuu2cPgCwEab09He+aM/SbF73inmkRWuGJkfEVPjBgugZ77D99h73sViJwDcdYRF1VMXWbCsgds4CH8CdxzHSSk+gTuO46QUn8Adx3FSyjUncDObNbNvmtnLZvaSmf1yLz5pZk+a2eu9PwdWpnccx3FuPcOImG0AvxJCeNbMygCeMbMnAfx7AH8XQvicmT0O4HEAv3qtnfXLEm2RqrUj6he2hDKp0sYCQAQhLAkXXUcIjCr9a6PGzj4lgPanHwWAJNbfkaWySEMphJymqEGpRDrlVFxfZ+GsWuVUnpNTd8g2VtbWOSj6R6WOVe7MunBIKi9lVzgxW8KpBwAtkU5WCcztAY7YQQQzqvvYafO1iCKV+piveaWjHYTtwOJsPmExe2qS045ePM8uwHNz7NhMRO3MzQ2R0nVdXG8AmYj7867jxyhWrwvRO+L7dlTUeZ08yOdcq7Fwm2yySxsAcgUe/03R56UxPvZUgQXQxjKn0H3l5N9TrJiwAJrL6bF6ZUm4QNfYJWsR3zuDuOYTeAjhUgjh2d7fNwC8AuAogI8D+GJvsy8C+Nmhj+o4juPcNNf1DtzMjgN4N4CnAMxsq0x/GQBnO3ccx3FuG0NP4GZWAvCXAD4bQtjxu1bYKo0jy+OY2WNm9rSZPb1Z4XW0jrNf2D5Wm+I1nuPsN4aawM0sg63J+09CCH/VC18xs8O9/z8MgF/wAAghPBFCOBFCOFEqDmdUcZy9YPtYzYqyaI6z3xhmFYoB+DyAV0IIv7vtv74K4FO9v38KwFduffMcx3GcQQzzmPF+AL8A4EUze74X+3UAnwPw52b2aQBnAfyba+0ojmOM9uVFXhYWYXT4bUxH2LAzIm84oFeXbFZ5tUKjyysAglhBEHXZitxVqzFEruzyiLb+ZlWe7w6fY1asYskm3J7+FAWAXg2hXmOFMGA1j6jEq77xVWHqlkhJsClWxZjYYyxWkSjLPQB0RAHj0OTrGou881fDYLBkZzsaYgVPHPPxR8SKiKkDbJkH9IoZtQonk/B2mxtsuT71+usUm5jkVRYHxnnVbwc6XUFGjMFjs7wKZUkU+r68xKtiRifLFHvnPZzOYaPLY+jyRT4/ALBR3memIFbfVLiNS/O84uTKGy/xsd94jmIzJb5WZ07ziiEAGJvgtCHrG1xkeUxntpBccwIPIfwjdN58APjI8IdyHMdxbiXuxHQcx0kpPoE7juOkFJ/AHcdxUsqurpUKAWj0FcrNl9hSns+yYDMvcltHQtgEgEjl/hai41qLRaDQ5c9O5VjwsQyLYkUhqh4/ov1Ni2t87Moai2SjRRbElB53/72cQ/rtDzxAsZbIq72xLIRkALV1FliyIle3iTQFWXENVTHmihDilIiXzQ+wUIvjZIVVG0I0vipmSPoExkKJ8zkL7RhlIaipnPIA0BRir0o5MD7ORaKPH+cUCFMTLFh2RRqBcpEFtVaiRUy1/Lcb+HySPIuGSns+Msn9eNc4729diMZos3UdANYXLlEsWxBjps7jrTr3FsU6SzzfZEXahnqb+2x5YUW2MVvgsVosi7oCA66Dwp/AHcdxUopP4I7jOCnFJ3DHcZyU4hO44zhOStlVETOKDMX8TrFi9g4W31rCGVitcq7i2hLnQwaA9TZ/L603WbyoBxaQprIsFplw+7VF6q6D0yxYmig8CwBBCHoXKizaVCu8XXmMhd/Dh9jpd3CCt3vzNBdRvTeEmMwAAA5YSURBVHxGu9tmxllAjYzPZ3rmEMUyWd4uEQ7SWMRWFkSR2RoLZACwJsZKXjghJ6e5jVcjjmMUyjuFvmMiB3ZLFO0tlYTo19FeuEadBatanfeZz/FYVf2ZEWKtibEaCXVxeoYLFQPA6CiPo7nLIrf1CovhTSHyHT7A+xtp8Gerq6sUm2jrhHj5jngWrYh7r8mfL43xZ+sX2YVcEbn5pye54POgPDrNFt/fuXEev8FcxHQcx/lnj0/gjuM4KcUncMdxnJTiE7jjOE5K2VURM5uNMHt8p4DRDOzKqldY0Dh2x90UOyMEIADorrAI1mmzWGBN4SbLsnsriMK1OZEGNCPETpUuFABmhPC3sCSKyoqUmnFgYXPuzBmKXTx3jmML3LfHjx6RbbzvjocptrnJhWvjmPunWmEhLhIOM+VEVIWKR3J6qGbEI0hTOPhWFmW9kcGYIelzeRbL7F5st8U4EFbZtSUW0QHAwNsePXKUYqffepNizRb3UyLEyalxIXAf4JSnnQFpheeX2JH73AucWnXuCrsXx8fYlTo+ziLm0gV2Q66vsaMxP8KfBYBClueRrUJhO2kLl++qKsBdYQE1K/r2jqMsbI+WWUQHgK7xvZPJ8T1RF2LnIPwJ3HEcJ6X4BO44jpNSfAJ3HMdJKcPUxJw1s2+a2ctm9pKZ/XIv/htmNmdmz/d+Pnb7m+s4juN8n2FEzDaAXwkhPGtmZQDPmNmTvf/7vRDCbw97sFqtjZde3OmejPLsjBoTKVTLZRYpZmaPy+M0muyiUilKY5GOtinqKZaFs64o3H6rayxCRiLVKgA0Gnzsao1FqUKRU6PmRJ3NpMX7U/Uiy6JviyUWmgAgzrD78fDRwxRTQm1dnIvqW6EPI5vl4zbFNQWAXI6PXSjyWMGA63A12n2i3rm5C7RNeZTHRr7L7T8/xwIfAJQL3PdTU+zuUzGVobZV5T5uixTCK8ssEAYhRgPAy6dOcexVrhnZETVmp4/weJk5wsJfWOfamVHMYn1+SqdnTiIWmGPj+2R5lRc4LFaFMC/u72yd77G8uBe7XRZFt2AhOyMctpUmOz4HMUxNzEsALvX+vmFmrwBgmdxxHMfZVa7rscTMjgN4N4CneqHPmNkLZvYFM+OqB47jOM5tY+gJ3MxKAP4SwGdDCOsA/gDAPQAextYT+u8M+NxjZva0mT29sTn8+kbH2W22j9V6Y/hfYx1nrxhqAjezDLYm7z8JIfwVAIQQroQQOiGELoA/BPCI+mwI4YkQwokQwomyKEnlOPuF7WM1r8qyOc4+45rvwM3MAHwewCshhN/dFj/cez8OAD8H4OS19lWr1/DiSzs3yyQsPpx4hL8LIiGwBOFIBIDSGNcFXFxYoJhKE6tqLBZHhLNKuNY6LRZxYpXLE0Bk3Pb3/Sifd0ec46ZwFU4dYmddq83C3wPjnDL00ff/mGzjpTMsXqnzjkzUIBV90VFCq+gfVdPSBjxrqN5tCtdtFOn6qYMICCR8t0Q60cV5FuETYQ+1oAXCQp5F0Pl5vr5KzC4IQbrS5PESCaV47sJ5ik1M8xgCgC54n3XhllWpgY8cY8GyUGY3pdDLkR9ngbdd4PkCAFo17vNKla/5sXveRrGOqOV57hKLvFkxXao6qUksHNUAkhz348gIX5vNQRqo2ucQ27wfwC8AeNHMnu/Ffh3AJ83sYWzdQ2cA/OLwh3Ucx3FulmFWofwjIMq8A1+/9c1xHMdxhsWdmI7jOCnFJ3DHcZyUsqvpZEMIaPXVyAuixt3CMrulRsdZLKjWdDrZUolFkmye08SuraxRTL0r6nRYFGuLWn9tIdKFASk6x0dZoPngT3E2gpUVTuX5zLe+SbEx4VTNCkH2wAx7sIoj3DcAkESi5qKyTgqXayR6MiPEyYxooxKsVTrgQcSiLmSno52cg+h2uqht7Fz2Gnf4+mZUPlvRHwXRJgDIiu7cEI7e+XlWtmKRtnYkz8Jmtc73WE6kRq2ucaphAIgTFt9GRb3VUoFjnSaLnWdFatzRPI8DU6mYhWAOACuLfJ+cO83H+fDkByj29js5nfLGvXdQrCHq094xy87QOMMCKAA0xRDoxJcoZtHwY9WfwB3HcVKKT+CO4zgpxSdwx3GclOITuOM4TkrZVREzigylPldjYtyEK1f4xf6IEGzyOWHfApAR8YKoZ1ircG4WUUYPlQqLQMqRmBH5PTOJrokZCw/h+TNcFzC02f13YJLzhiXCBZcXwlCzxgLZmVe/p9soXJImRCQlEI6IY2eEo7Uj+rsh8pB0hDAIDHCBijTB10u328Hm5s70n0lbOenE9c3yc1Ehpx2EeSEgJyJFaXOJncRtMYaW11mY79R4/B4os4jebWlncyxcw/fecxfFxkpce3NEpPutb3Ba1cqycDGL+rTxCIuiAFDdEAsSxDiorrLA2K1zG49Ms5t7fZXbmBeLIyzWKUMaon5quyXE/gF1dBX+BO44jpNSfAJ3HMdJKT6BO47jpBSfwB3HcVKKT+CO4zgpZVdXoSAA7T5LdCdihbs6P8efFTb1O+88Lg/TETb3sih0W7iT7bIVsTKlIYrCBrEqIhF26VJRf0fGwkK+tsiFb2tCsR8Vtvk44eOo/MxJzG3M57Q9OY5420aTr0M2Kyzy4thdYa9X6QdUCgCdVR3odnj8qOLJcXx9zyrdbqDisiWxyqgp8rVHorGtAXnhL62wfd1EW2ORs1qNwY1FLg6cFUurGiJf+6C0DyoP+kSZiyyPl8coVizw2FArRiJ1bDHWGk3+LAAEcT4zk5zfvCaSbc+dP0uxepNX7lQ2xSokcX5xovtx7uIqxfKzIg1FjnPED8KfwB3HcVKKT+CO4zgp5ZoTuJnlzew7ZvY9M3vJzP5zL36XmT1lZm+Y2Z+ZmXbVOI7jOLeFYZ7AGwB+PITwELYq0P+0mT0K4LcA/F4I4W0AVgB8+vY103Ecx+lnmJJqAcBm75+Z3k8A8OMAfr4X/yKA3wDwB1fbV7fbQLP+xo7Y3CK/sM8LC223JUQgka8aAGaPHqJYnPAvCOUy22VVLvEr85cpVhX2+ki0RwmbABCJ786RAn/eOsJOLARCZXHPZkRxYCGmZUTfbCGKPouc3qHLx1ZCogkRUMmnkchVXSyIwtIAOkLErFZYqOoOKIB9Pagi0RYpMZvPqiaEQACYn2fhOpvl63H8ruMUyxj305EZHvsZ1cldUai4sik2BEwJ10qTFXnzK2sswgdhcY/F2KhWOO1DNHCsMgWRc19lZMjleGw12ryYoSmEzfl5Fo0Tkc8eAC5f4n2WstwXk4eGP8eh3oGbWdwraDwP4EkAbwJYDSF8f0RfAMCVAhzHcZzbxlATeAihE0J4GMAxAI8AeGDYA5jZY2b2tJk9XavffJIhx7ldbB+r6mnbcfYb17UKJYSwCuCbAH4EwLjZD1IJHgMgFm8DIYQnQggnQggnVIY6x9kvbB+rmQGvvhxnPzHMKpSDZjbe+/sIgI8CeAVbE/m/7m32KQBfuV2NdBzHcZhhHjMOA/iimcXYmvD/PITwNTN7GcCfmtlvAngOwOevtaOMNXEws/NBffRt7IY8e4FftSQRuw9XRY5kACgVWbyYmOTYepVFhaIoClsULs6GyLFcb7B4NpbVrqqucJ5FMQuW+RH+vHIVRqLYsCrGnAjHptpu0HEA3rYlRCklVCknZpxhwUa9vsjmdOFlZeBTfavEzmsR9wl1HeHIrda5rQXj8ZIIoQwAiiXhqhXi5OYqC3qJuD79+fYBIJdhgVDp/ypHPQC0hBs4I1y6tSYLtReFy/HAJC8eKBd5nCvhud3kewwAOl3lNuV7vlPlfOIZ4SRWU6Mqll0TBaMRtIiZz7JTdW2ZzyfKDv/6bphVKC8AeLeIn8bW+3DHcRxnD3AnpuM4TkrxCdxxHCel+ATuOI6TUiyoKr6362BmCwC+r2ocALC4awe/vfi57E+udS53hhAOqv/wsZoK/iWdixyruzqB7ziw2dMhhBN7cvBbjJ/L/uRWnYv3yf7Ez8VfoTiO46QWn8Adx3FSyl5O4E/s4bFvNX4u+5NbdS7eJ/uTf/HnsmfvwB3HcZybw1+hOI7jpJRdn8DN7KfN7FSvFNvju338m8XMvmBm82Z2clts0syeNLPXe39O7GUbh8HMZs3sm2b2cq9U3i/34qk7F+D2lP7zsbo/8LF6FUIIu/YDIMZWMYi7AWQBfA/Ag7vZhltwDh8E8B4AJ7fF/iuAx3t/fxzAb+11O4c4j8MA3tP7exnAawAeTOO59NpqAEq9v2cAPAXgUQB/DuATvfh/B/BLQ+7Px+o++fGxepV97XLDfwTA32z7968B+LW97tAbOI/jfTfFKQCHe38/DODUXrfxBs7pK9hKFfzP4VwKAJ4F8D5smSOSXnzH+LvGPnys7tMfH6v//2e3X6EcBXB+27//uZRimwkhXOr9/TKAmb1szPViZsexlXHyKaT4XG5x6T8fq/sQH6s7cRHzFhO2vj5Ts7THzEoA/hLAZ0MIO5JOp+1cwk2U/vuXSNqur49VZrcn8DkAs9v+PbAUW8q4YmaHAaD35/wet2cozCyDrRviT0IIf9ULp/JcthNuoPSfwMfqPsLHqma3J/DvAri3p7ZmAXwCwFd3uQ23g69iq6wckJLycmZm2Kqi9EoI4Xe3/VfqzgW4LaX/fKzuE3ysXoU9eGn/MWypyG8C+A97LSLcQPu/BOASgBa23lN9GsAUgL8D8DqA/w1gcq/bOcR5fABbv3K+AOD53s/H0nguvfN5F7ZK+70A4CSA/9iL3w3gOwDeAPAXAHLXsU8fq/vgx8fq4B93YjqO46QUFzEdx3FSik/gjuM4KcUncMdxnJTiE7jjOE5K8QnccRwnpfgE7jiOk1J8Anccx0kpPoE7juOklP8HdD2IPUPDBfkAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 288
        },
        "id": "6v3KoYaLn7c-",
        "outputId": "ee026bd6-31ac-415a-d15f-4392ccdb9477"
      },
      "source": [
        "import pandas as pd\n",
        "import seaborn as sn\n",
        "\n",
        "def get_conf_matrix(x_test, y_test):\n",
        "    m = np.zeros((10,10))\n",
        "    acc = 0\n",
        "    i = 0\n",
        "    for x in x_test:      \n",
        "        yp = np.argmax(model.predict(np.expand_dims(x, axis=0)))    \n",
        "        m[np.argmax(y_test[i]),yp] += 1\n",
        "        i = i+1\n",
        "    m = m/len(y_test)               \n",
        "\n",
        "    return m\n",
        "\n",
        "m = get_conf_matrix(x_test[:1000], y_test[:1000])\n",
        "df_cm = pd.DataFrame(m, range(10), range(10))\n",
        "sn.set(font_scale=1.2) # for label size\n",
        "sn.heatmap(df_cm, annot=True, annot_kws={\"size\": 8})"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7ff936f71590>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c78JHZpLn_0N"
      },
      "source": [
        "## Создание и обучение тестовой модели №2"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "whys-F83oG8d",
        "outputId": "ac6a6aeb-030d-484b-b460-7f3733593b25"
      },
      "source": [
        "batch_size = 64\n",
        "epochs = 200\n",
        "\n",
        "model = Sequential()\n",
        "model.add(Conv2D(64, kernel_size=3, activation='relu', input_shape=x_train.shape[1:]))\n",
        "model.add(Conv2D(32, kernel_size=3, activation='relu'))\n",
        "model.add(Flatten())\n",
        "model.add(Dense(num_classes, activation='softmax'))\n",
        "\n",
        "opt = keras.optimizers.Adam()\n",
        "\n",
        "model.compile(loss='categorical_crossentropy',\n",
        "              optimizer=opt,\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "model.summary()"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d_2 (Conv2D)            (None, 30, 30, 64)        1792      \n",
            "_________________________________________________________________\n",
            "conv2d_3 (Conv2D)            (None, 28, 28, 32)        18464     \n",
            "_________________________________________________________________\n",
            "flatten_1 (Flatten)          (None, 25088)             0         \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 10)                250890    \n",
            "=================================================================\n",
            "Total params: 271,146\n",
            "Trainable params: 271,146\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 653
        },
        "id": "fbcaAIiYewpl",
        "outputId": "353160dc-a473-402c-dc6e-b00901823c81"
      },
      "source": [
        "print('Not using data augmentation.')\n",
        "\n",
        "model.fit(x_train, y_train,\n",
        "          batch_size=batch_size,\n",
        "          epochs=epochs,\n",
        "          validation_data=(x_test, y_test),\n",
        "          shuffle=True,\n",
        "          callbacks=[plot]\n",
        "          )"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 22/200\n",
            "334/782 [===========>..................] - ETA: 1:27 - loss: 0.0643 - accuracy: 0.9787"
          ],
          "name": "stdout"
        },
        {
          "output_type": "error",
          "ename": "KeyboardInterrupt",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-9-85c896553618>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      6\u001b[0m           \u001b[0mvalidation_data\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_test\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m           \u001b[0mshuffle\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m           \u001b[0mcallbacks\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mplot\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      9\u001b[0m           )\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/engine/training.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, validation_batch_size, validation_freq, max_queue_size, workers, use_multiprocessing)\u001b[0m\n\u001b[1;32m   1181\u001b[0m                 _r=1):\n\u001b[1;32m   1182\u001b[0m               \u001b[0mcallbacks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mon_train_batch_begin\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstep\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1183\u001b[0;31m               \u001b[0mtmp_logs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain_function\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0miterator\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1184\u001b[0m               \u001b[0;32mif\u001b[0m \u001b[0mdata_handler\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshould_sync\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1185\u001b[0m                 \u001b[0mcontext\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0masync_wait\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/def_function.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwds)\u001b[0m\n\u001b[1;32m    887\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    888\u001b[0m       \u001b[0;32mwith\u001b[0m \u001b[0mOptionalXlaContext\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_jit_compile\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 889\u001b[0;31m         \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwds\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    890\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    891\u001b[0m       \u001b[0mnew_tracing_count\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexperimental_get_tracing_count\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/def_function.py\u001b[0m in \u001b[0;36m_call\u001b[0;34m(self, *args, **kwds)\u001b[0m\n\u001b[1;32m    915\u001b[0m       \u001b[0;31m# In this case we have created variables on the first call, so we run the\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    916\u001b[0m       \u001b[0;31m# defunned version which is guaranteed to never create variables.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 917\u001b[0;31m       \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_stateless_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwds\u001b[0m\u001b[0;34m)\u001b[0m  \u001b[0;31m# pylint: disable=not-callable\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    918\u001b[0m     \u001b[0;32melif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_stateful_fn\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    919\u001b[0m       \u001b[0;31m# Release the lock early so that multiple threads can perform the call\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   3022\u001b[0m        filtered_flat_args) = self._maybe_define_function(args, kwargs)\n\u001b[1;32m   3023\u001b[0m     return graph_function._call_flat(\n\u001b[0;32m-> 3024\u001b[0;31m         filtered_flat_args, captured_inputs=graph_function.captured_inputs)  # pylint: disable=protected-access\n\u001b[0m\u001b[1;32m   3025\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   3026\u001b[0m   \u001b[0;34m@\u001b[0m\u001b[0mproperty\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36m_call_flat\u001b[0;34m(self, args, captured_inputs, cancellation_manager)\u001b[0m\n\u001b[1;32m   1959\u001b[0m       \u001b[0;31m# No tape is watching; skip to running the function.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1960\u001b[0m       return self._build_call_outputs(self._inference_function.call(\n\u001b[0;32m-> 1961\u001b[0;31m           ctx, args, cancellation_manager=cancellation_manager))\n\u001b[0m\u001b[1;32m   1962\u001b[0m     forward_backward = self._select_forward_and_backward_functions(\n\u001b[1;32m   1963\u001b[0m         \u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36mcall\u001b[0;34m(self, ctx, args, cancellation_manager)\u001b[0m\n\u001b[1;32m    594\u001b[0m               \u001b[0minputs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    595\u001b[0m               \u001b[0mattrs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mattrs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 596\u001b[0;31m               ctx=ctx)\n\u001b[0m\u001b[1;32m    597\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    598\u001b[0m           outputs = execute.execute_with_cancellation(\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/execute.py\u001b[0m in \u001b[0;36mquick_execute\u001b[0;34m(op_name, num_outputs, inputs, attrs, ctx, name)\u001b[0m\n\u001b[1;32m     58\u001b[0m     \u001b[0mctx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mensure_initialized\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     59\u001b[0m     tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,\n\u001b[0;32m---> 60\u001b[0;31m                                         inputs, attrs, num_outputs)\n\u001b[0m\u001b[1;32m     61\u001b[0m   \u001b[0;32mexcept\u001b[0m \u001b[0mcore\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_NotOkStatusException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     62\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mname\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "D3HQ3ZcwrERB"
      },
      "source": [
        "## Создание и обучение тестовой модели №3"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NRaVOzZErEiu",
        "outputId": "36918049-5787-40c4-bbd8-0b4f790928b0"
      },
      "source": [
        "batch_size = 64\n",
        "epochs = 250\n",
        "\n",
        "model = Sequential()\n",
        "\n",
        "model.add(Conv2D(32, kernel_size=3, activation='relu', kernel_initializer='he_normal', input_shape=x_train.shape[1:])) \n",
        "model.add(Flatten())\n",
        "model.add(Dropout(0.25)) \n",
        "model.add(Dense(128, activation='relu')) \n",
        "model.add(Dense(num_classes, activation='softmax'))\n",
        "\n",
        "opt = keras.optimizers.Adam()\n",
        "\n",
        "model.compile(loss='categorical_crossentropy',\n",
        "              optimizer=opt,\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "model.summary()"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d_4 (Conv2D)            (None, 30, 30, 32)        896       \n",
            "_________________________________________________________________\n",
            "flatten_2 (Flatten)          (None, 28800)             0         \n",
            "_________________________________________________________________\n",
            "dropout_1 (Dropout)          (None, 28800)             0         \n",
            "_________________________________________________________________\n",
            "dense_3 (Dense)              (None, 128)               3686528   \n",
            "_________________________________________________________________\n",
            "dense_4 (Dense)              (None, 10)                1290      \n",
            "=================================================================\n",
            "Total params: 3,688,714\n",
            "Trainable params: 3,688,714\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 653
        },
        "id": "_YAIb-2yrEua",
        "outputId": "805dcc2d-d113-4358-8bd0-700ca8ffb32c"
      },
      "source": [
        "print('Not using data augmentation.')\n",
        "\n",
        "model.fit(x_train, y_train,\n",
        "          batch_size=batch_size,\n",
        "          epochs=epochs,\n",
        "          validation_data=(x_test, y_test),\n",
        "          shuffle=True,\n",
        "          callbacks=[plot]\n",
        "          )"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 15/250\n",
            "196/782 [======>.......................] - ETA: 44s - loss: 0.3326 - accuracy: 0.8846"
          ],
          "name": "stdout"
        },
        {
          "output_type": "error",
          "ename": "KeyboardInterrupt",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-11-85c896553618>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      6\u001b[0m           \u001b[0mvalidation_data\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_test\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m           \u001b[0mshuffle\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m           \u001b[0mcallbacks\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mplot\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      9\u001b[0m           )\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/engine/training.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, validation_batch_size, validation_freq, max_queue_size, workers, use_multiprocessing)\u001b[0m\n\u001b[1;32m   1181\u001b[0m                 _r=1):\n\u001b[1;32m   1182\u001b[0m               \u001b[0mcallbacks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mon_train_batch_begin\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstep\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1183\u001b[0;31m               \u001b[0mtmp_logs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain_function\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0miterator\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1184\u001b[0m               \u001b[0;32mif\u001b[0m \u001b[0mdata_handler\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshould_sync\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1185\u001b[0m                 \u001b[0mcontext\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0masync_wait\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/def_function.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwds)\u001b[0m\n\u001b[1;32m    887\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    888\u001b[0m       \u001b[0;32mwith\u001b[0m \u001b[0mOptionalXlaContext\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_jit_compile\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 889\u001b[0;31m         \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwds\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    890\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    891\u001b[0m       \u001b[0mnew_tracing_count\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexperimental_get_tracing_count\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/def_function.py\u001b[0m in \u001b[0;36m_call\u001b[0;34m(self, *args, **kwds)\u001b[0m\n\u001b[1;32m    915\u001b[0m       \u001b[0;31m# In this case we have created variables on the first call, so we run the\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    916\u001b[0m       \u001b[0;31m# defunned version which is guaranteed to never create variables.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 917\u001b[0;31m       \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_stateless_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwds\u001b[0m\u001b[0;34m)\u001b[0m  \u001b[0;31m# pylint: disable=not-callable\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    918\u001b[0m     \u001b[0;32melif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_stateful_fn\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    919\u001b[0m       \u001b[0;31m# Release the lock early so that multiple threads can perform the call\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   3022\u001b[0m        filtered_flat_args) = self._maybe_define_function(args, kwargs)\n\u001b[1;32m   3023\u001b[0m     return graph_function._call_flat(\n\u001b[0;32m-> 3024\u001b[0;31m         filtered_flat_args, captured_inputs=graph_function.captured_inputs)  # pylint: disable=protected-access\n\u001b[0m\u001b[1;32m   3025\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   3026\u001b[0m   \u001b[0;34m@\u001b[0m\u001b[0mproperty\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36m_call_flat\u001b[0;34m(self, args, captured_inputs, cancellation_manager)\u001b[0m\n\u001b[1;32m   1959\u001b[0m       \u001b[0;31m# No tape is watching; skip to running the function.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1960\u001b[0m       return self._build_call_outputs(self._inference_function.call(\n\u001b[0;32m-> 1961\u001b[0;31m           ctx, args, cancellation_manager=cancellation_manager))\n\u001b[0m\u001b[1;32m   1962\u001b[0m     forward_backward = self._select_forward_and_backward_functions(\n\u001b[1;32m   1963\u001b[0m         \u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36mcall\u001b[0;34m(self, ctx, args, cancellation_manager)\u001b[0m\n\u001b[1;32m    594\u001b[0m               \u001b[0minputs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    595\u001b[0m               \u001b[0mattrs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mattrs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 596\u001b[0;31m               ctx=ctx)\n\u001b[0m\u001b[1;32m    597\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    598\u001b[0m           outputs = execute.execute_with_cancellation(\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/execute.py\u001b[0m in \u001b[0;36mquick_execute\u001b[0;34m(op_name, num_outputs, inputs, attrs, ctx, name)\u001b[0m\n\u001b[1;32m     58\u001b[0m     \u001b[0mctx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mensure_initialized\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     59\u001b[0m     tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,\n\u001b[0;32m---> 60\u001b[0;31m                                         inputs, attrs, num_outputs)\n\u001b[0m\u001b[1;32m     61\u001b[0m   \u001b[0;32mexcept\u001b[0m \u001b[0mcore\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_NotOkStatusException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     62\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mname\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "18WWPkgHqdx9"
      },
      "source": [
        "## Создание и обучение тестовой модели №4"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5bEuxTgVqkmw",
        "outputId": "2cfad7f2-ebbd-466a-9caf-376ce2de9e0c"
      },
      "source": [
        "batch_size = 32\n",
        "epochs = 150\n",
        "\n",
        "model = keras.Sequential([\n",
        "    Conv2D(32, (3, 3), padding='same', activation='relu', kernel_initializer='he_normal', input_shape=x_train.shape[1:]),\n",
        "    MaxPooling2D((2,2), strides=2),\n",
        "    Conv2D(64, (3,3), padding='same', activation='relu'),\n",
        "    MaxPooling2D(2, strides=2),\n",
        "    Flatten(),\n",
        "    Dense(128, activation='relu'),\n",
        "    Dense(num_classes, kernel_initializer='he_normal', activation='softmax')\n",
        "])\n",
        "\n",
        "opt = keras.optimizers.Adam()\n",
        "\n",
        "model.compile(loss='categorical_crossentropy',\n",
        "              optimizer=opt,\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "model.summary()\n",
        "\n",
        "\n",
        "# model.add(Dense(32, kernel_initializer='he_normal'))\n"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_4\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d_8 (Conv2D)            (None, 32, 32, 32)        896       \n",
            "_________________________________________________________________\n",
            "max_pooling2d_5 (MaxPooling2 (None, 16, 16, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_9 (Conv2D)            (None, 16, 16, 64)        18496     \n",
            "_________________________________________________________________\n",
            "max_pooling2d_6 (MaxPooling2 (None, 8, 8, 64)          0         \n",
            "_________________________________________________________________\n",
            "flatten_4 (Flatten)          (None, 4096)              0         \n",
            "_________________________________________________________________\n",
            "dense_7 (Dense)              (None, 128)               524416    \n",
            "_________________________________________________________________\n",
            "dense_8 (Dense)              (None, 10)                1290      \n",
            "=================================================================\n",
            "Total params: 545,098\n",
            "Trainable params: 545,098\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 653
        },
        "id": "lsRDAvsbqp2w",
        "outputId": "610b9c7c-d259-44b5-cde5-c962a89846e8"
      },
      "source": [
        "print('Not using data augmentation.')\n",
        "\n",
        "model.fit(x_train, y_train,\n",
        "          batch_size=batch_size,\n",
        "          epochs=epochs,\n",
        "          validation_data=(x_test, y_test),\n",
        "          shuffle=True,\n",
        "          callbacks=[plot]\n",
        "          )"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 14/150\n",
            " 476/1563 [========>.....................] - ETA: 52s - loss: 0.0568 - accuracy: 0.9810"
          ],
          "name": "stdout"
        },
        {
          "output_type": "error",
          "ename": "KeyboardInterrupt",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-17-85c896553618>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      6\u001b[0m           \u001b[0mvalidation_data\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_test\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m           \u001b[0mshuffle\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m           \u001b[0mcallbacks\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mplot\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      9\u001b[0m           )\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/engine/training.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, validation_batch_size, validation_freq, max_queue_size, workers, use_multiprocessing)\u001b[0m\n\u001b[1;32m   1181\u001b[0m                 _r=1):\n\u001b[1;32m   1182\u001b[0m               \u001b[0mcallbacks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mon_train_batch_begin\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstep\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1183\u001b[0;31m               \u001b[0mtmp_logs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain_function\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0miterator\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1184\u001b[0m               \u001b[0;32mif\u001b[0m \u001b[0mdata_handler\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshould_sync\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1185\u001b[0m                 \u001b[0mcontext\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0masync_wait\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/def_function.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwds)\u001b[0m\n\u001b[1;32m    887\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    888\u001b[0m       \u001b[0;32mwith\u001b[0m \u001b[0mOptionalXlaContext\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_jit_compile\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 889\u001b[0;31m         \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwds\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    890\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    891\u001b[0m       \u001b[0mnew_tracing_count\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexperimental_get_tracing_count\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/def_function.py\u001b[0m in \u001b[0;36m_call\u001b[0;34m(self, *args, **kwds)\u001b[0m\n\u001b[1;32m    915\u001b[0m       \u001b[0;31m# In this case we have created variables on the first call, so we run the\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    916\u001b[0m       \u001b[0;31m# defunned version which is guaranteed to never create variables.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 917\u001b[0;31m       \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_stateless_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwds\u001b[0m\u001b[0;34m)\u001b[0m  \u001b[0;31m# pylint: disable=not-callable\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    918\u001b[0m     \u001b[0;32melif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_stateful_fn\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    919\u001b[0m       \u001b[0;31m# Release the lock early so that multiple threads can perform the call\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   3022\u001b[0m        filtered_flat_args) = self._maybe_define_function(args, kwargs)\n\u001b[1;32m   3023\u001b[0m     return graph_function._call_flat(\n\u001b[0;32m-> 3024\u001b[0;31m         filtered_flat_args, captured_inputs=graph_function.captured_inputs)  # pylint: disable=protected-access\n\u001b[0m\u001b[1;32m   3025\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   3026\u001b[0m   \u001b[0;34m@\u001b[0m\u001b[0mproperty\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36m_call_flat\u001b[0;34m(self, args, captured_inputs, cancellation_manager)\u001b[0m\n\u001b[1;32m   1959\u001b[0m       \u001b[0;31m# No tape is watching; skip to running the function.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1960\u001b[0m       return self._build_call_outputs(self._inference_function.call(\n\u001b[0;32m-> 1961\u001b[0;31m           ctx, args, cancellation_manager=cancellation_manager))\n\u001b[0m\u001b[1;32m   1962\u001b[0m     forward_backward = self._select_forward_and_backward_functions(\n\u001b[1;32m   1963\u001b[0m         \u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36mcall\u001b[0;34m(self, ctx, args, cancellation_manager)\u001b[0m\n\u001b[1;32m    594\u001b[0m               \u001b[0minputs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    595\u001b[0m               \u001b[0mattrs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mattrs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 596\u001b[0;31m               ctx=ctx)\n\u001b[0m\u001b[1;32m    597\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    598\u001b[0m           outputs = execute.execute_with_cancellation(\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/execute.py\u001b[0m in \u001b[0;36mquick_execute\u001b[0;34m(op_name, num_outputs, inputs, attrs, ctx, name)\u001b[0m\n\u001b[1;32m     58\u001b[0m     \u001b[0mctx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mensure_initialized\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     59\u001b[0m     tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,\n\u001b[0;32m---> 60\u001b[0;31m                                         inputs, attrs, num_outputs)\n\u001b[0m\u001b[1;32m     61\u001b[0m   \u001b[0;32mexcept\u001b[0m \u001b[0mcore\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_NotOkStatusException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     62\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mname\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0o47sbut73nw"
      },
      "source": [
        "## Создание и обучение тестовой модели №5"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lIk7y40W76eH",
        "outputId": "6dae5378-b397-4bd4-f5d6-074e5ed99d8a"
      },
      "source": [
        "batch_size = 64\n",
        "epochs = 250\n",
        "\n",
        "model = Sequential()\n",
        "model.add(Conv2D(32, (3,3), input_shape=x_train.shape[1:], kernel_initializer='he_normal'))\n",
        "model.add(LeakyReLU(0.1))\n",
        "model.add(BatchNormalization())\n",
        "model.add(MaxPooling2D(2))\n",
        "\n",
        "model.add(Conv2D(64, (3,3)))\n",
        "model.add(LeakyReLU(0.1))\n",
        "model.add(BatchNormalization())\n",
        "model.add(MaxPooling2D(2))\n",
        "\n",
        "model.add(Flatten())\n",
        "model.add(Dropout(0.3))\n",
        "\n",
        "model.add(Dense(32, kernel_initializer='he_normal'))\n",
        "model.add(LeakyReLU(0.1))\n",
        "\n",
        "model.add(Dense(num_classes, kernel_initializer='he_normal'))\n",
        "model.add(Activation('softmax'))\n",
        "\n",
        "opt = keras.optimizers.Adam()\n",
        "\n",
        "model.compile(loss='categorical_crossentropy',\n",
        "              optimizer=opt,\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "model.summary()"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_5\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d_10 (Conv2D)           (None, 30, 30, 32)        896       \n",
            "_________________________________________________________________\n",
            "leaky_re_lu_3 (LeakyReLU)    (None, 30, 30, 32)        0         \n",
            "_________________________________________________________________\n",
            "batch_normalization_2 (Batch (None, 30, 30, 32)        128       \n",
            "_________________________________________________________________\n",
            "max_pooling2d_7 (MaxPooling2 (None, 15, 15, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_11 (Conv2D)           (None, 13, 13, 64)        18496     \n",
            "_________________________________________________________________\n",
            "leaky_re_lu_4 (LeakyReLU)    (None, 13, 13, 64)        0         \n",
            "_________________________________________________________________\n",
            "batch_normalization_3 (Batch (None, 13, 13, 64)        256       \n",
            "_________________________________________________________________\n",
            "max_pooling2d_8 (MaxPooling2 (None, 6, 6, 64)          0         \n",
            "_________________________________________________________________\n",
            "flatten_5 (Flatten)          (None, 2304)              0         \n",
            "_________________________________________________________________\n",
            "dropout_2 (Dropout)          (None, 2304)              0         \n",
            "_________________________________________________________________\n",
            "dense_9 (Dense)              (None, 32)                73760     \n",
            "_________________________________________________________________\n",
            "leaky_re_lu_5 (LeakyReLU)    (None, 32)                0         \n",
            "_________________________________________________________________\n",
            "dense_10 (Dense)             (None, 10)                330       \n",
            "_________________________________________________________________\n",
            "activation_1 (Activation)    (None, 10)                0         \n",
            "=================================================================\n",
            "Total params: 93,866\n",
            "Trainable params: 93,674\n",
            "Non-trainable params: 192\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 653
        },
        "id": "2Ip9lolT7_Xa",
        "outputId": "d1c0d712-ff48-4a10-a78d-31ed35a6755e"
      },
      "source": [
        "print('Not using data augmentation.')\n",
        "\n",
        "model.fit(x_train, y_train,\n",
        "          batch_size=batch_size,\n",
        "          epochs=epochs,\n",
        "          validation_data=(x_test, y_test),\n",
        "          shuffle=True,\n",
        "          callbacks=[plot]\n",
        "          )"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 42/250\n",
            "105/782 [===>..........................] - ETA: 1:05 - loss: 0.4037 - accuracy: 0.8564"
          ],
          "name": "stdout"
        },
        {
          "output_type": "error",
          "ename": "KeyboardInterrupt",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-19-85c896553618>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      6\u001b[0m           \u001b[0mvalidation_data\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_test\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m           \u001b[0mshuffle\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m           \u001b[0mcallbacks\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mplot\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      9\u001b[0m           )\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/engine/training.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, validation_batch_size, validation_freq, max_queue_size, workers, use_multiprocessing)\u001b[0m\n\u001b[1;32m   1181\u001b[0m                 _r=1):\n\u001b[1;32m   1182\u001b[0m               \u001b[0mcallbacks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mon_train_batch_begin\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstep\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1183\u001b[0;31m               \u001b[0mtmp_logs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain_function\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0miterator\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1184\u001b[0m               \u001b[0;32mif\u001b[0m \u001b[0mdata_handler\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshould_sync\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1185\u001b[0m                 \u001b[0mcontext\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0masync_wait\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/def_function.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwds)\u001b[0m\n\u001b[1;32m    887\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    888\u001b[0m       \u001b[0;32mwith\u001b[0m \u001b[0mOptionalXlaContext\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_jit_compile\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 889\u001b[0;31m         \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwds\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    890\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    891\u001b[0m       \u001b[0mnew_tracing_count\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexperimental_get_tracing_count\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/def_function.py\u001b[0m in \u001b[0;36m_call\u001b[0;34m(self, *args, **kwds)\u001b[0m\n\u001b[1;32m    915\u001b[0m       \u001b[0;31m# In this case we have created variables on the first call, so we run the\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    916\u001b[0m       \u001b[0;31m# defunned version which is guaranteed to never create variables.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 917\u001b[0;31m       \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_stateless_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwds\u001b[0m\u001b[0;34m)\u001b[0m  \u001b[0;31m# pylint: disable=not-callable\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    918\u001b[0m     \u001b[0;32melif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_stateful_fn\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    919\u001b[0m       \u001b[0;31m# Release the lock early so that multiple threads can perform the call\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   3022\u001b[0m        filtered_flat_args) = self._maybe_define_function(args, kwargs)\n\u001b[1;32m   3023\u001b[0m     return graph_function._call_flat(\n\u001b[0;32m-> 3024\u001b[0;31m         filtered_flat_args, captured_inputs=graph_function.captured_inputs)  # pylint: disable=protected-access\n\u001b[0m\u001b[1;32m   3025\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   3026\u001b[0m   \u001b[0;34m@\u001b[0m\u001b[0mproperty\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36m_call_flat\u001b[0;34m(self, args, captured_inputs, cancellation_manager)\u001b[0m\n\u001b[1;32m   1959\u001b[0m       \u001b[0;31m# No tape is watching; skip to running the function.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1960\u001b[0m       return self._build_call_outputs(self._inference_function.call(\n\u001b[0;32m-> 1961\u001b[0;31m           ctx, args, cancellation_manager=cancellation_manager))\n\u001b[0m\u001b[1;32m   1962\u001b[0m     forward_backward = self._select_forward_and_backward_functions(\n\u001b[1;32m   1963\u001b[0m         \u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36mcall\u001b[0;34m(self, ctx, args, cancellation_manager)\u001b[0m\n\u001b[1;32m    594\u001b[0m               \u001b[0minputs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    595\u001b[0m               \u001b[0mattrs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mattrs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 596\u001b[0;31m               ctx=ctx)\n\u001b[0m\u001b[1;32m    597\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    598\u001b[0m           outputs = execute.execute_with_cancellation(\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/execute.py\u001b[0m in \u001b[0;36mquick_execute\u001b[0;34m(op_name, num_outputs, inputs, attrs, ctx, name)\u001b[0m\n\u001b[1;32m     58\u001b[0m     \u001b[0mctx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mensure_initialized\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     59\u001b[0m     tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,\n\u001b[0;32m---> 60\u001b[0;31m                                         inputs, attrs, num_outputs)\n\u001b[0m\u001b[1;32m     61\u001b[0m   \u001b[0;32mexcept\u001b[0m \u001b[0mcore\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_NotOkStatusException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     62\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mname\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "POp_T49xFAQo"
      },
      "source": [
        "## Создание и обучение тестовой модели №6"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ehPTTtWNFDbX",
        "outputId": "e815e136-55de-48e3-c822-e9c6e39c5f27"
      },
      "source": [
        "batch_size = 32\n",
        "epochs = 300\n",
        "\n",
        "model = Sequential()\n",
        "model.add(Conv2D(64, (3,3), input_shape=x_train.shape[1:], kernel_initializer='he_normal'))\n",
        "model.add(LeakyReLU(0.4))\n",
        "model.add(BatchNormalization())\n",
        "model.add(MaxPooling2D(2))\n",
        "\n",
        "model.add(Conv2D(128, (3,3)))\n",
        "model.add(LeakyReLU(0.2))\n",
        "model.add(BatchNormalization())\n",
        "model.add(MaxPooling2D(2))\n",
        "\n",
        "model.add(Flatten())\n",
        "model.add(Dropout(0.3))\n",
        "\n",
        "model.add(Dense(64, kernel_initializer='he_normal'))\n",
        "model.add(LeakyReLU(0.4))\n",
        "\n",
        "model.add(Dense(num_classes, kernel_initializer='he_normal'))\n",
        "model.add(Activation('softmax'))\n",
        "\n",
        "opt = keras.optimizers.Adam()\n",
        "\n",
        "model.compile(loss='categorical_crossentropy',\n",
        "              optimizer=opt,\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "model.summary()"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_6\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d_12 (Conv2D)           (None, 30, 30, 64)        1792      \n",
            "_________________________________________________________________\n",
            "leaky_re_lu_6 (LeakyReLU)    (None, 30, 30, 64)        0         \n",
            "_________________________________________________________________\n",
            "batch_normalization_4 (Batch (None, 30, 30, 64)        256       \n",
            "_________________________________________________________________\n",
            "max_pooling2d_9 (MaxPooling2 (None, 15, 15, 64)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_13 (Conv2D)           (None, 13, 13, 128)       73856     \n",
            "_________________________________________________________________\n",
            "leaky_re_lu_7 (LeakyReLU)    (None, 13, 13, 128)       0         \n",
            "_________________________________________________________________\n",
            "batch_normalization_5 (Batch (None, 13, 13, 128)       512       \n",
            "_________________________________________________________________\n",
            "max_pooling2d_10 (MaxPooling (None, 6, 6, 128)         0         \n",
            "_________________________________________________________________\n",
            "flatten_6 (Flatten)          (None, 4608)              0         \n",
            "_________________________________________________________________\n",
            "dropout_3 (Dropout)          (None, 4608)              0         \n",
            "_________________________________________________________________\n",
            "dense_11 (Dense)             (None, 64)                294976    \n",
            "_________________________________________________________________\n",
            "leaky_re_lu_8 (LeakyReLU)    (None, 64)                0         \n",
            "_________________________________________________________________\n",
            "dense_12 (Dense)             (None, 10)                650       \n",
            "_________________________________________________________________\n",
            "activation_2 (Activation)    (None, 10)                0         \n",
            "=================================================================\n",
            "Total params: 372,042\n",
            "Trainable params: 371,658\n",
            "Non-trainable params: 384\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 653
        },
        "id": "YQtIzqINJKvH",
        "outputId": "483811b8-b1c0-45b3-faae-c6988fe9c508"
      },
      "source": [
        "print('Not using data augmentation.')\n",
        "\n",
        "model.fit(x_train, y_train,\n",
        "          batch_size=batch_size,\n",
        "          epochs=epochs,\n",
        "          validation_data=(x_test, y_test),\n",
        "          shuffle=True,\n",
        "          callbacks=[plot]\n",
        "          )"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 20/300\n",
            "  84/1563 [>.............................] - ETA: 2:40 - loss: 0.3963 - accuracy: 0.8624"
          ],
          "name": "stdout"
        },
        {
          "output_type": "error",
          "ename": "KeyboardInterrupt",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-21-85c896553618>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      6\u001b[0m           \u001b[0mvalidation_data\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_test\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m           \u001b[0mshuffle\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m           \u001b[0mcallbacks\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mplot\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      9\u001b[0m           )\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/engine/training.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, validation_batch_size, validation_freq, max_queue_size, workers, use_multiprocessing)\u001b[0m\n\u001b[1;32m   1181\u001b[0m                 _r=1):\n\u001b[1;32m   1182\u001b[0m               \u001b[0mcallbacks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mon_train_batch_begin\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstep\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1183\u001b[0;31m               \u001b[0mtmp_logs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain_function\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0miterator\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1184\u001b[0m               \u001b[0;32mif\u001b[0m \u001b[0mdata_handler\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshould_sync\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1185\u001b[0m                 \u001b[0mcontext\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0masync_wait\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/def_function.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwds)\u001b[0m\n\u001b[1;32m    887\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    888\u001b[0m       \u001b[0;32mwith\u001b[0m \u001b[0mOptionalXlaContext\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_jit_compile\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 889\u001b[0;31m         \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwds\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    890\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    891\u001b[0m       \u001b[0mnew_tracing_count\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexperimental_get_tracing_count\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/def_function.py\u001b[0m in \u001b[0;36m_call\u001b[0;34m(self, *args, **kwds)\u001b[0m\n\u001b[1;32m    915\u001b[0m       \u001b[0;31m# In this case we have created variables on the first call, so we run the\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    916\u001b[0m       \u001b[0;31m# defunned version which is guaranteed to never create variables.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 917\u001b[0;31m       \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_stateless_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwds\u001b[0m\u001b[0;34m)\u001b[0m  \u001b[0;31m# pylint: disable=not-callable\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    918\u001b[0m     \u001b[0;32melif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_stateful_fn\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    919\u001b[0m       \u001b[0;31m# Release the lock early so that multiple threads can perform the call\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   3022\u001b[0m        filtered_flat_args) = self._maybe_define_function(args, kwargs)\n\u001b[1;32m   3023\u001b[0m     return graph_function._call_flat(\n\u001b[0;32m-> 3024\u001b[0;31m         filtered_flat_args, captured_inputs=graph_function.captured_inputs)  # pylint: disable=protected-access\n\u001b[0m\u001b[1;32m   3025\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   3026\u001b[0m   \u001b[0;34m@\u001b[0m\u001b[0mproperty\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36m_call_flat\u001b[0;34m(self, args, captured_inputs, cancellation_manager)\u001b[0m\n\u001b[1;32m   1959\u001b[0m       \u001b[0;31m# No tape is watching; skip to running the function.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1960\u001b[0m       return self._build_call_outputs(self._inference_function.call(\n\u001b[0;32m-> 1961\u001b[0;31m           ctx, args, cancellation_manager=cancellation_manager))\n\u001b[0m\u001b[1;32m   1962\u001b[0m     forward_backward = self._select_forward_and_backward_functions(\n\u001b[1;32m   1963\u001b[0m         \u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36mcall\u001b[0;34m(self, ctx, args, cancellation_manager)\u001b[0m\n\u001b[1;32m    594\u001b[0m               \u001b[0minputs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    595\u001b[0m               \u001b[0mattrs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mattrs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 596\u001b[0;31m               ctx=ctx)\n\u001b[0m\u001b[1;32m    597\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    598\u001b[0m           outputs = execute.execute_with_cancellation(\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/execute.py\u001b[0m in \u001b[0;36mquick_execute\u001b[0;34m(op_name, num_outputs, inputs, attrs, ctx, name)\u001b[0m\n\u001b[1;32m     58\u001b[0m     \u001b[0mctx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mensure_initialized\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     59\u001b[0m     tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,\n\u001b[0;32m---> 60\u001b[0;31m                                         inputs, attrs, num_outputs)\n\u001b[0m\u001b[1;32m     61\u001b[0m   \u001b[0;32mexcept\u001b[0m \u001b[0mcore\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_NotOkStatusException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     62\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mname\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
          ]
        }
      ]
    }
  ]
}